# Era — Full Content > Complete content from era.app for LLM context. Generated at build time. ## Homepage Era is an AI-native personal finance platform. The homepage introduces how Era connects your bank accounts to any AI assistant through the Model Context Protocol (MCP). It showcases the core value proposition: one connection, every AI agent. Key sections include a hero with the tagline about connecting AI to your finances, a visual demonstration of the MCP integration flow, a grid of supported AI clients (Claude, ChatGPT, Gemini, Cursor, and more), feature highlights covering spending analysis, cash flow forecasting, automated categorization, and custom rules. The page includes a FAQ section addressing common questions about security, supported institutions, pricing, and how MCP works. Social proof and trust indicators emphasize bank-level encryption, SOC 2 compliance, and the fact that Era never stores bank credentials directly. ## Context MCP server documentation The Context MCP server documentation at https://era.app/help/mcp-server-era-context explains how MCP-compatible clients connect to Era Context. The production server is named Era Context, uses Streamable HTTP at https://context.era.app, also accepts the explicit https://context.era.app/mcp endpoint, is protected by OAuth, and publishes MCP directory metadata at https://era.app/.well-known/mcp/server-card.json. The public tool catalog is generated from Era's live MCP server metadata, so it stays aligned with the tools an assistant can discover and call. Current catalog summary: 62 production-registered tools and up to 48 tools available to ordinary assistants. Billing mutations require explicit mcp:billing-write consent, and the public catalog focuses on tools intended for normal user-facing discovery. ## Pricing Era offers four pricing tiers: - **See** (Free): Read-only MCP tools, connect up to 2 financial accounts, 1 personal view, basic spending categorization. Designed for users who want to try AI-powered finance with zero commitment. - **Organize** ($9.99/month): Full rules engine, MCP read and write tools, unlimited accounts, custom categories, transaction tagging, and knowledge memory. For users who want AI to actively help organize their finances. - **Automate** ($24.99/month): Everything in Organize plus automation rules, MCP write tools, AI-driven financial management, and priority support. For users who want hands-off financial management. - **Optimize** ($39.99/month): Everything in Automate plus Era brokerage with model portfolios, autonomous optimization via Agency (Era's first-party AI), unlimited transfers, and priority support. For users who want AI to actively grow their wealth. ## Company The Company page introduces the team behind Era. Built by a team with backgrounds at Stripe, Robinhood, CashApp, Apple, Google, SoFi, and Coursera. The page outlines Era's mission: to make finance work for everyone by removing the barriers between people, their money, and the AI tools that can help them make better decisions. It covers the company's approach to security (bank-level encryption, SOC 2 compliance, regulated data partners), its commitment to the open MCP standard rather than proprietary integrations, and the vision for agent-native finance where AI assistants become the primary interface for managing money. ## Articles ### 5 things your AI remembers that your budgeting app doesn't Budgeting apps have gotten smarter. Monarch has a built-in AI assistant. Copilot offers AI-powered insights. Even legacy tools are adding chat interfaces. But there's a fundamental difference between an AI that lives *inside* your budgeting app and one that actually *remembers you*. Here are five things Era's cross-agent memory stores that your budgeting app's AI can't — and why each one matters. ## 1. Your savings goal, in every AI you use Tell Monarch's AI you're saving for a house down payment. That goal lives in Monarch. Open ChatGPT to run a scenario? No idea you're saving for anything. Ask Claude to review your spending? Blank slate. You start over every time. With Era, you tell one AI your goal and every connected AI inherits it. Tell Claude "I'm saving $1,200 a month and want $40,000 by March." Open ChatGPT the next day and ask about your spending — it already knows what you're working toward. Switch to Gemini for a different question the following week. Same context, same goal, no re-explaining. Built-in AI is useful as long as you stay in the app. Cross-agent memory is useful everywhere. ## 2. Your spending preferences — the ones that always get overridden Every budgeting app categorizes your gym membership as "health and fitness." Every app suggests cutting it when you're over budget. You override it manually, every month, because it's essential to you — not a luxury. With Era, you tell any connected AI "my gym membership is essential spending, not discretionary." That preference gets stored in your memory profile. Every AI, every conversation, every analysis respects it from that point on. No more monthly argument with your own budgeting tool about whether $50 is a luxury. This applies to any preference that doesn't fit the app's default categories: a monthly donation you consider non-negotiable, a business expense buried in a personal account, a shared bill that only one of you pays but both of you track. ## 3. Your household arrangement Money is rarely just one person. Most households have a division of who pays what, but budgeting apps don't have a field for "my partner covers the mortgage, I cover utilities." They just show you transactions and let you draw your own conclusions. Era's AI can hold that context. "My partner and I split rent 60/40, they pay directly, I reimburse by bank transfer on the 1st." One sentence, stored once, available to every AI you use. Now when you ask any agent for a spending breakdown, it already understands your household structure. No more "why is my rent so low?" — the AI already knows. This is particularly useful when you're analyzing cash flow or running "can I afford this?" scenarios. The AI has the full picture, not just your half of it. ## 4. Notes to your future self This one's underrated. Budgeting apps are good at tracking what happened. They're not built for "remind me that I negotiated a better rate on my car insurance — it renews in September." Era lets you store notes that any AI can surface when they're relevant. "I negotiated my gym membership down from $75 to $50 in April — worth trying again at renewal." "My brokerage bonus vests in October — factor that into any projections." "I've been meaning to cancel the annual subscription I stopped using — it auto-renews in November." These aren't transactions. They're not categories. They're just things you know about your own financial life that you'd otherwise have to remember yourself, write in a notes app, or re-explain every time you start a new conversation. ## 5. The plan you're building across multiple conversations Most financial planning happens in fragments. You discuss your debt paydown strategy with Claude on a Tuesday. You explore refinancing options with ChatGPT on a Thursday. You check your progress with Gemini two weeks later. If each conversation starts fresh, you're never building on anything — you're just running one-off queries. With cross-agent memory, each conversation picks up where the last one left off. Your debt paydown plan from Tuesday is still in context on Thursday. Your refinancing research is available two weeks later. You're building a coherent picture over time, not starting from scratch every session. Monarch, Copilot, and YNAB each store conversation history within their own app. But that history stays siloed — it doesn't travel with you to Claude or ChatGPT or Cursor. Era's memory layer sits beneath all of them, which means continuity doesn't depend on staying inside one product. ## The comparison at a glance | | Era cross-agent memory | Monarch / Copilot built-in AI | |---|---|---| | Memory lives in | Your Era Context — you own it | The app's system | | Accessible by | Any MCP-compatible AI | That app's AI only | | Persists across AI clients | Yes | No | | Goals travel with you | Yes | No | | Works with Claude | Yes | No | | Works with ChatGPT | Yes | No | | Memory limit (free plan) | 50 facts | N/A (single app only) | At time of writing, no budgeting app offers cross-agent memory. Era is the only personal finance platform where your financial context is portable — not locked to a single product or a single AI. ## What this means in practice The practical difference is this: with a built-in AI assistant, you get smart analysis as long as you stay in the app. With Era's cross-agent memory, you get smart analysis wherever you're working — in Claude, in ChatGPT, in Cursor, in whatever comes next. Your financial context belongs to you, not to the app where you first typed it. For a deeper look at how Era compares to Monarch, Copilot, and YNAB across other dimensions — pricing, platform, automation, and more — [Era vs. Monarch vs. Copilot vs. YNAB: 2026 comparison](/articles/era-vs-monarch-vs-copilot-vs-ynab) covers the full picture. And if you're curious about how memory actually works across agents under the hood, [can AI agents share financial memory?](/articles/can-ai-agents-share-financial-memory) explains the mechanics. Ready to try it? Era's free plan gets you started with two connected accounts and up to 50 stored memory facts — no credit card required. ### Best AI for remembering your finances in 2026 The difference between an AI that helps you with your finances and one that actually *knows your finances* comes down to memory. Not conversation history — that's temporary. Not chat summaries — those stay locked in one tool. Actual persistent, portable memory that follows you from one AI to the next. In 2026, a handful of platforms are competing in this space. Here's an honest look at what each offers for personal finance memory specifically. ## What we're comparing This isn't a general AI assistant comparison or a budgeting app shootout. We're focused on one specific question: which platform is best at remembering your financial context across AI tools and over time? The four contenders: Era Context, Mem0, MemoryLake, and the built-in memory features of generic AI assistants like Claude and ChatGPT. ## Era Context **What it is:** A personal finance platform built from the ground up for AI-native access via MCP (the Model Context Protocol). The memory system — called Era Context — stores your financial goals, preferences, household context, and notes in a shared profile that any connected AI can read. **How memory works:** When you tell any connected AI something about your finances, it's stored in your Era Context profile. Connect Claude, ChatGPT, Cursor, or any other MCP-compatible client, and each one reads from the same profile. Your goals, preferences, and context are available everywhere — not siloed in one tool. **Finance-specific depth:** This is Era's key advantage. The memory system is purpose-built for personal finance. It understands the difference between a stated goal ("I'm saving for a house"), a spending preference ("gym membership is essential"), and a household arrangement ("my partner covers rent"). These distinctions shape how every connected AI answers your financial questions. **Memory limits:** Free plan holds up to 50 facts — enough for a solid core profile. The Organize plan expands to 200, and Automate and above are unlimited. See [pricing](/pricing) for full details. **Cross-agent portability:** Full. Any MCP-compatible AI reads from and writes to the same profile. Retract a fact and it disappears across all clients simultaneously. **Best for:** People who already use AI tools throughout their day and want their finances to be part of that workflow. If you have preferences about which AI model you use, Era is the only option that lets you bring your own AI while keeping financial context portable. ## Mem0 **What it is:** An open-source AI memory infrastructure layer aimed primarily at developers. Mem0 provides APIs for storing and retrieving user memories across conversations, and powers the memory features in several third-party AI applications. **How memory works:** Mem0 extracts entities, preferences, and facts from conversation text and stores them in a vector database. Retrieval is semantic — you query by relevance, not by browsing a list. Developers integrate Mem0 into their own applications. **Finance-specific depth:** Essentially none. Mem0 is domain-agnostic — it stores whatever you tell it. There's no built-in understanding of financial concepts, no integration with bank accounts, no transaction data, and no structure around goals vs. preferences vs. context. You'd need to build all of that yourself. **Cross-agent portability:** Depends entirely on the implementation. Mem0 is infrastructure, not a consumer product. Whether a Mem0-backed app gives you portable, multi-client memory depends on what the app developer built. **Best for:** Developers building AI applications who want memory infrastructure they control. Not a consumer-ready personal finance memory product at time of writing. ## MemoryLake **What it is:** A newer entrant in the AI memory infrastructure space, targeting enterprise and developer use cases. MemoryLake focuses on long-term, structured memory for AI agents, with an emphasis on searchability and memory versioning. **How memory works:** MemoryLake maintains a structured memory store with version history. Memory entries can be tagged, searched, and rolled back. It supports multi-agent architectures where several AI agents read from a shared memory pool. **Finance-specific depth:** Like Mem0, MemoryLake is domain-agnostic. It provides memory infrastructure; you build the financial layer on top. At time of writing, there are no finance-specific integrations, no bank account connections, and no consumer product — it's a developer platform. **Cross-agent portability:** Strong at the infrastructure level, but again, it's what you build with it. A developer could build a portable finance memory product using MemoryLake, but that product doesn't exist out of the box. **Best for:** Development teams building multi-agent AI systems who need durable, versioned memory infrastructure. Not a direct competitor for personal finance users. ## Built-in memory in Claude and ChatGPT **What it is:** Both Claude and ChatGPT now offer built-in memory features that let them remember things across conversations within their respective platforms. **How memory works:** You can ask Claude to remember something, and it'll recall it in future Claude conversations. Same with ChatGPT. These memories live inside each platform's system and are not accessible outside of it. **Finance-specific depth:** Moderate. Claude and ChatGPT have strong financial reasoning capabilities, and their built-in memory means they can remember your goals over time — within the same tool. But the memory is not structured specifically for finance; it's the same memory system used for everything else. **Cross-agent portability:** None. Claude's memory doesn't travel to ChatGPT. ChatGPT's memory doesn't travel to Claude. If you use both — or if you switch — you start over. **Best for:** People who use one AI tool consistently and want it to remember their preferences over time. If you never switch between AI tools, built-in memory may be sufficient. ## The comparison | | Era Context | Mem0 | MemoryLake | Claude/ChatGPT built-in | |---|---|---|---|---| | Consumer-ready product | Yes | Developer only | Developer only | Yes | | Finance-specific design | Yes | No | No | No | | Bank account integration | Yes | No | No | No | | Cross-agent portability | Full | Depends on implementation | Depends on implementation | None | | Works with Claude | Yes | Depends | Depends | N/A | | Works with ChatGPT | Yes | Depends | Depends | N/A | | Free plan available | Yes | Open-source (self-host) | N/A | Yes (limited) | | Memory structured by type | Yes (goals, preferences, context) | No | Partially | No | ## The verdict for personal finance If you're looking for something that works today, without building anything, and that genuinely understands the difference between a financial goal and a spending preference, Era Context is the clearest choice in 2026. It's the only consumer-ready platform that combines bank account integration with cross-agent memory that actually travels with you. Mem0 and MemoryLake are strong developer tools, but they're infrastructure — you'd need to build the financial layer yourself. Claude and ChatGPT's built-in memory is useful, but it doesn't cross platforms. The broader question isn't which of these is best in a vacuum — it's which one fits how you actually work. If you use multiple AI tools and want your financial context to be consistent across all of them, Era is currently the only option that solves that problem out of the box. For a look at how Era's memory stacks up against the built-in AI features of Monarch, Copilot, and YNAB specifically, [5 things your AI remembers that your budgeting app doesn't](/articles/5-things-your-ai-remembers-that-your-budgeting-app-doesnt) goes into that comparison in detail. And if you want to get started today, [how to set up cross-agent memory for your finances](/articles/how-to-set-up-cross-agent-memory) walks through the setup step by step. ### How to set up cross-agent memory for your finances You told Claude about your savings goal last week. Now you open ChatGPT and have to explain it all over again. That's the problem cross-agent memory solves — and setting it up takes about five minutes. This guide walks you through connecting Era Context to your AI of choice so your financial context follows you everywhere, automatically. ## What you'll need - An Era account (free to start — no credit card required) - A bank account at a supported institution - An MCP-compatible AI client: Claude, ChatGPT, OpenClaw, Cursor, or Gemini all work If you already use Claude for financial questions and want to understand what memory actually stores, [what Era Context stores and how to manage it](/articles/what-era-context-stores-and-how-to-manage-it) covers that in depth. If you just want to get set up now, keep reading. ## Step 1: Create your Era account and connect a bank Go to [era.app](https://era.app) and sign up. The Basic plan is free and includes two connected accounts and up to 50 stored memory facts — enough to get a feel for everything before you commit to anything. Once you're in Era Context, tap the connect flow and link a bank account. Era connects through MX, a regulated financial data provider, so your bank credentials are never stored by Era. You authenticate directly with your bank, complete any two-factor step your bank requires, and your accounts appear in Era Context within about 30 seconds. If you have checking and savings at the same bank, both show up automatically. ## Step 2: Add Era Context to your AI client Every MCP-compatible client follows the same basic pattern: add Era Context's MCP server URL and complete an authorization step. Here's how it looks for the most common clients. ### Claude Desktop Open Claude Desktop, go to **Settings → MCP Servers**, and add a new server: ```json { "mcpServers": { "era-context": { "url": "https://context.era.app" } } } ``` Save, and Claude will prompt you to authorize the connection. Once you confirm, Claude has live access to your financial data and your memory profile. {{connect-claude-button}} ### ChatGPT In ChatGPT's settings, look for the MCP or external connections section and add `https://context.era.app` as the server URL. Complete the OAuth consent screen that appears. ### Other MCP-compatible clients The URL is always `https://context.era.app`. Any client that supports MCP follows the same pattern: paste the URL, complete the authorization, and the connection is live. ## Step 3: Start building your memory profile Here's the part that surprises most people: you don't need to set up memory explicitly. You just talk to your AI like you normally would, and Era stores what matters. Try it. Open a conversation with your connected AI and mention something real: - *"I'm saving $800 a month toward a house down payment — I want to be ready by next spring."* - *"My gym membership is essential spending, not discretionary — please don't flag it as something to cut."* - *"My partner covers the mortgage, I cover utilities and groceries."* Each of these gets stored in your Era Context memory profile. Now open a different AI client and ask about your finances — it already knows, without you repeating yourself. On the Basic plan, your memory profile holds up to 50 facts — enough for a solid financial context across goals, preferences, and household arrangements. The Organize plan expands that to 200 facts, and Automate and above remove the limit entirely. See [pricing](/pricing) if you're curious about what's included at each tier. ## Step 4: Check what's stored on your Knowledge page Era's Knowledge page (at [/app/knowledge](/app/knowledge)) shows everything currently in your memory profile: your stated goals, saved preferences, facts your AI has learned from your conversations, and any inferences it's made that you've confirmed. From there you can: - **Review** what's stored so far - **Edit** any fact that's outdated or wrong - **Retract** anything you want removed — it disappears across all connected agents instantly You're always in control of what your AI remembers. If something in the list no longer applies — say, a savings goal you've already hit — you can remove it with one tap, and it's gone everywhere. ## Step 5: Connect additional AI clients One Era account works with as many AI clients as you want. Connect Claude for in-depth analysis, ChatGPT for quick questions, Cursor for financial scripting while you work — each one reads from and writes to the same memory profile. When you connect a new client, it inherits your full existing memory profile immediately. No setup, no re-explaining your situation from scratch. The new assistant already knows your goals, your preferences, and your context. To connect another client, repeat Step 2. Each connection requires its own authorization step — you control which clients have access, independently. ## What cross-agent memory actually feels like The moment that makes it click is usually the first time you switch clients mid-task. You start a conversation with Claude about whether you can afford a vacation without derailing your savings goal. Claude knows your goal because you mentioned it two weeks ago. You switch to ChatGPT to run a quick scenario. It also knows your goal — same Era Context, different client. For a deeper look at why this portability matters and how it compares to what built-in AI assistants offer, [why your financial AI memory should be portable](/articles/why-financial-ai-memory-should-be-portable) walks through the tradeoffs clearly. And if you want to understand the full picture of what gets stored and how sharing works across agents, [can AI agents share financial memory?](/articles/can-ai-agents-share-financial-memory) covers the mechanics. ## FAQs **How long does setup take?** About five minutes from account creation to having an AI client with live bank access and memory working. Longer if your bank has a complicated two-factor flow, but rarely more than ten minutes. **Does memory work on the free plan?** Yes. The Basic plan includes up to 50 stored memory facts with full read and write access — you can review, edit, and retract any stored fact. Paid plans add higher limits: Organize gives you 200 facts, and Automate and above remove the cap entirely. **What if I want to start fresh?** Head to the Knowledge page and retract whatever you want cleared. There's no "reset everything" button (to protect against accidents), but removing individual facts is fast, and each removal is instant across all connected agents. **Which AI clients work with Era?** Any MCP-compatible client — including Claude, ChatGPT, OpenClaw, Cursor, Manus, and Gemini. When a new client adds MCP support, it works with Era immediately. **Is my memory private?** Your memory profile is private to you. It's never shared with other users, never used to train AI models, and only accessible to the AI clients you've explicitly authorized. You can revoke any client's access at any time from Era Context settings. ### What Era Context stores and how to manage it One of the most common questions people ask after connecting an AI to their finances is: "What exactly does it remember about me?" It's a fair question — and the answer should be completely visible to you. This article covers exactly what Era Context stores in your memory profile, where you can see it, and how to change or remove anything you don't want stored. ## What goes into your memory profile Your memory profile is a collection of facts about your financial life. Some are things you've stated directly. Others are things your AI has inferred from your transactions or conversations, which you've since confirmed. Facts come in a few forms: **Stated goals.** Things you've explicitly told your AI you're working toward. "I want to save $15,000 for a house down payment by next spring." "I'm trying to pay off my credit card before the intro APR expires." These are stored as-is, in plain language. **Preferences.** How you want your AI to interpret your situation. "My gym membership is essential, not discretionary." "My partner covers the mortgage — it's not part of my solo budget." These shape how your AI frames analysis and answers questions. **Household and financial context.** Standing facts about your situation that help AI give better answers. Who covers which bills. How income is structured. What accounts are shared. What recurring charges exist for specific reasons. **Confirmed inferences.** When your AI notices a pattern — say, a regular transfer to a savings account every payday — it may surface an inference like "you appear to be saving roughly $500 per month automatically." If you confirm it, that inference becomes a stored fact. **Notes.** Things you've asked your AI to remember for later. Contract renewal dates, negotiated rates, plans you want to revisit. Era's AI never stores raw transaction data as memory facts. What your transactions look like is accessible separately through the transaction tools. Memory is for things about *you* and *your intentions*, not a copy of your bank feed. ## The Knowledge page: what you'll see Visit [/app/knowledge](/app/knowledge) in Era Context to see your full memory profile. The page shows every stored fact: its content, the date it was created or last updated, and which AI client added it. The display is intentionally readable — these are your own words and your AI's summaries, not internal field labels or technical identifiers. If something looks unfamiliar, it's because an AI inferred it from your behavior and added it after you confirmed; the confirmation event is recorded. ## How to view your memory in detail The Knowledge page lets you browse your stored facts and filter them by type or date. For each fact, you can see: - **The fact itself** — what was stored, in plain language - **When it was created** — so you can tell if it's recent or something from months ago - **How it was added** — whether you stated it directly, confirmed an inference, or asked your AI to note it - **Which agent added it** — so you know whether Claude, ChatGPT, or another client contributed it Nothing is hidden. If your AI knows it, it appears here. ## How to edit a stored fact Facts go stale. Goals get hit. Arrangements change. When a stored fact no longer reflects reality, you can update it directly on the Knowledge page. Tap any fact to open it, then edit the text. The update is immediate and applies across all connected agents — the next time any AI accesses your memory profile, it sees the updated version. If you've hit your savings goal and the stored target is now wrong, update it. If your household arrangement changed, update that too. Your memory profile should reflect your current life, not a snapshot from six months ago. ## How to retract (remove) a stored fact If you want a fact completely removed, retract it. Retraction is permanent and immediate: the fact is deleted from your Era Context and every connected agent loses access to it instantly. There's no undo. Retracted facts are gone. This is intentional — memory without real deletion isn't memory you can trust. To retract, tap the fact on the Knowledge page and choose the retract option. You can also ask any connected AI directly: "Forget that I was considering switching jobs." The AI will handle the retraction through the same mechanism. After retraction, the fact doesn't exist anywhere in your profile — not in Claude's context, not in ChatGPT's, not in any client you've connected. ## Memory limits by plan How many facts your profile can hold depends on your plan: - **Basic (free):** 50 facts — enough for core goals, key preferences, and household context - **Organize:** 200 facts — room for detailed context, multiple goals, and richer notes - **Automate and above:** No limit — your AI can store as much context as your situation needs You can always retract facts you no longer need to keep your profile tidy, or upgrade to a higher plan if your situation calls for more room. See [pricing](/pricing) for what's included at each tier. ## Privacy: who else can see your memory Your memory profile is private to you. It is never shared with other Era users, never visible to Era employees in daily operations, and never used to train AI models. Each AI client you connect can read from your profile only for the queries you initiate. No client can access your memory without your explicit authorization, and you can revoke any client's access at any time from Era Context settings. For a detailed look at how Era's security model works end to end — encryption, OAuth consent, revocation, and what never leaves Era's systems — [is it safe to give AI your financial data?](/articles/is-it-safe-to-give-ai-your-financial-data) covers all of it. ## How memory relates to cross-agent sharing Because your memory profile is stored in Era Context and not inside any individual AI client, every connected agent reads from the same source. This is what makes cross-agent memory work: it's not that Claude and ChatGPT are somehow synced — it's that they both read from your Era Context profile independently. When you retract a fact, all agents lose access simultaneously because all agents were reading from the same place. There's no sync step, no delay, no residual copy in a specific client's history. For more on how cross-agent sharing works and what it means for your financial context, [can AI agents share financial memory?](/articles/can-ai-agents-share-financial-memory) and [why your financial AI memory should be portable](/articles/why-financial-ai-memory-should-be-portable) both go into the underlying architecture in plain language. ## FAQs **Can I export my memory profile?** Your stored facts are visible on the Knowledge page, and you can copy or screenshot them. A full export tool is on the roadmap. **What happens to my memory if I downgrade my plan?** Existing facts are preserved on downgrade — nothing already stored is deleted automatically. If you want to stay within a lower plan's included limit, you can retract facts you no longer need. **Does my memory profile sync in real time?** Yes. Updates, edits, and retractions propagate to all connected agents immediately. There's no batch sync or delay. **Can my AI add facts without asking me?** Your AI can add facts based on things you say in conversation. If you'd prefer to be prompted before anything is stored, you can ask your AI to request confirmation before adding anything new — it will follow that preference going forward. **What if the same fact is stored twice?** The Knowledge page will show both. You can retract the duplicate and keep the one you prefer. There's no automatic deduplication because slight variations in how a fact is phrased may both be intentional. ### Why your financial AI memory should be portable Your financial goals, preferences, and context should belong to you — not to the app or AI assistant where you first typed them. When memory is portable, every AI you use starts informed. When it is not, you spend every new conversation re-explaining yourself. ## What does "portable memory" mean for AI? Portable memory means your financial context — your goals, preferences, spending rules, and notes — lives in a layer you own, not inside any single assistant's conversation history. Most AI assistants store what you tell them locally. Claude remembers your preferences inside Claude. ChatGPT stores facts inside ChatGPT. That memory is useful, but it is siloed. Move to a different assistant and none of it comes with you. Era Context takes a different approach. It stores your financial context in a personal MCP server — a layer you own that sits beneath any AI you use. Connect Claude, ChatGPT, or any other MCP-compatible client to Era Context, and every one of them reads from the same source. For a full picture of how Era Context works, see [what is Era Context?](/articles/what-is-era-context) ## Why does built-in AI memory create lock-in? Apps like Monarch Money, Copilot Money, and Cleo embed AI directly into their product. The AI is a feature of the platform. Your financial data and your AI memory both live in that system. That is a deliberate design choice. But it has a consequence: your context is only as portable as the app allows. If the app changes its AI, your history does not transfer. If you want to use a different assistant — Claude for analysis, ChatGPT for planning, Cursor for scripting custom reports — you cannot bring your financial memory with you. You rebuild from scratch in every new context. | | Era Context | Built-in AI (Monarch, Copilot, Cleo) | |---|---|---| | Memory location | Your Era Context — you own it | The app's system | | Accessible by | Any MCP-compatible AI | That app's AI only | | Portable across assistants | Yes | No | | Works with Claude | Yes | No | | Works with ChatGPT | Yes | No | | Forget from everywhere | Yes | Depends on the app | | Revoke access per assistant | Yes | Account-level only | ## What happens when you switch AI assistants? Without portable memory, switching assistants means starting over. You re-explain your savings target, your household setup, your spending priorities. The new assistant has no context. Every conversation feels like the first one. With Era Context, switching is transparent. Your goals and preferences are stored in your Context profile. Connect a new assistant and it reads from the same source as every other assistant you use. Tell Claude you want to build a three-month emergency fund. Open ChatGPT the next day and ask whether you are on track. It already knows what you are working toward — not because you told it, but because both assistants read from the same memory layer. For help choosing which assistant suits which tasks, [how to pick an AI agent for your money](/articles/how-to-pick-ai-agent-for-your-money) walks through the tradeoffs. ## What does portable memory look like in practice? A few concrete examples: **Goal persistence.** Set a savings target in one conversation. Every assistant you connect respects it without re-entry. **Preference continuity.** Tell one assistant your gym membership is essential, not discretionary. Every subsequent spending review, across every assistant, reflects that classification. **Forget everywhere.** Ask any connected assistant to forget something and it is removed from your Era Context — not just from that conversation, but from every assistant's access at once. **New assistant, zero setup.** Connect a new MCP-compatible client tomorrow and it inherits your complete financial context immediately. This is what cross-agent memory means in practice. For a deeper look at how it works, [can AI agents share financial memory?](/articles/can-ai-agents-share-financial-memory) covers the details. ## Is portable memory secure? Portability does not mean less secure. Era Context encrypts your data with AES-256 at rest and TLS 1.3 in transit. Your bank credentials are never stored. Your financial data is never used to train AI models and never shared without your explicit permission. You control which AI clients connect. Revoke access to any one of them at any time. Disconnect an assistant and it loses access to your memory immediately. Your data stays with Era, not with any individual AI provider. For a full breakdown of how Era handles security, see [is it safe to give AI your financial data?](/articles/is-it-safe-to-give-ai-your-financial-data) ## How do you get started with Era Context? 1. Create an Era account and connect your bank accounts through Era Context. 2. Go to the Era Context settings page and copy your MCP server URL: `https://context.era.app` 3. Add Era Context to your AI client. For Claude, the fastest path is the [Claude Connectors directory](https://claude.com/connectors/era-context) — find Era Context there and connect in one click. For other clients, paste the MCP URL directly into their MCP server settings. 4. Repeat for any additional clients. Each one immediately reads from your shared Context. 5. Start a conversation. Tell your AI about a financial goal or preference. It is now available to every connected assistant. Era Context is free to start. See [pricing](/pricing) for plans that include the full rules engine and expanded account connections. For a complete step-by-step walkthrough, [how to set up cross-agent memory for your finances](/articles/how-to-set-up-cross-agent-memory) is the practical companion to this article. And if you're curious about what actually gets stored once you're up and running, [what Era Context stores and how to manage it](/articles/what-era-context-stores-and-how-to-manage-it) explains exactly what your memory profile contains and how to edit or remove anything in it. ## FAQs **What is portable financial memory?** Your financial goals and preferences stored in a layer any AI can read — not inside one assistant. When you switch tools, your context follows you automatically. **How is Era Context different from Claude's or ChatGPT's built-in memory?** Claude and ChatGPT each store memory locally, scoped to their own sessions. That memory does not cross between assistants. Era Context stores your financial context in a shared layer every connected assistant reads from and writes to, regardless of which one you used last. **Which AI assistants work with Era Context?** Any assistant that supports MCP — including Claude, ChatGPT, OpenClaw, Cursor, Gemini, and dozens more. When a new client adds MCP support, Era Context works with it immediately. **What happens to my memory if I stop using one AI assistant?** Nothing. Your goals, rules, and preferences are stored in Era Context, not in the assistant. Stop using one, add a new one, or switch entirely — your financial context stays intact. **Is my financial data secure when shared across assistants?** Each assistant accesses your Era Context through a controlled, revocable connection. You grant access per client and can remove it at any time. Data is encrypted in transit and at rest, and your bank credentials are never stored by Era. ### How to text your AI about your money with Poke and Era You can text a question like "what did I spend on groceries this month?" and get a real answer — pulled live from your connected accounts. Poke is an MCP-compatible messaging client from The Interactive Company. Connect it to Era Context once, and your finances — checking, savings, credit cards, and more — are a text message away. ## What is Poke? [Poke](https://poke.com/) is a messaging-first AI client built by The Interactive Company. Instead of opening a dedicated AI app, you send a message — the way you already communicate. Poke supports the Model Context Protocol (MCP), which means it can connect to external data sources like Era Context and pull real information into its responses. ## Why text your money? Most personal finance questions are quick. You don't need a full app session to answer "did my rent go through?" or "how much have I spent on coffee this month?" A text-message interaction strips away the friction. You ask, you get the answer, you move on. Era Context gives Poke 33 tools across 7 groups — accounts, transactions, insights, activity, rules, and more. Poke can call the right tool automatically based on what you ask. You never need to know which tool it uses. ## What you need - A [Poke](https://poke.com/) account - An Era account (the Basic tier is free) - About five minutes ## How to connect Era Context to Poke 1. **Create your Era account.** Sign up at [era.app](https://era.app). The Basic tier is free and includes two connected accounts with read-only MCP access. 2. **Connect your accounts.** From Era Context, start the connect flow. Era uses MX, a regulated financial data provider, to securely link your institution. You can connect checking accounts, savings accounts, credit cards, and investment accounts — all in one place. Your bank credentials are never stored by Era — MX handles the authentication handshake. 3. **Set up the Era integration in Poke.** Open [Poke](https://poke.com/), navigate to **Settings → Integrations**, and click **New template**. Add Era Context as the MCP server using this URL: ``` https://context.era.app ``` Save the template. That's the entire configuration. 4. **Complete the OAuth authorization.** When Poke connects for the first time, you'll be redirected to an authorization screen. Review the permissions and confirm. This explicit authorization step applies every time you connect a new AI client — nothing happens silently. 5. **Send your first message.** Try: "What are my current account balances?" Poke pulls live figures from all your connected accounts and replies. ## What can you ask over text? Once connected, you can ask Poke anything you could ask any Era-connected AI: - "How much did I spend this month compared to last month?" - "Show me all my recurring charges" - "What did I spend at restaurants this week?" - "What's my cash flow for the past 30 days?" - "Did any transactions come in over $200 today?" - "What are my biggest spending categories?" Poke picks the right Era tool for each question. You just send the message. ## Can you set up recurring financial updates? Yes. Poke supports recurring messages, so you can schedule financial check-ins the same way you'd set a reminder. Some useful recurring setups: - **Monday morning recap** — "Send me a weekly spending summary every Monday at 8am" - **End-of-month review** — "On the last day of each month, show me my top 5 spending categories" - **Daily balance check** — "Text me my checking and credit card balances every morning" - **Subscription watch** — "Every Sunday, remind me to check for any new recurring charges" Set these up in Poke once and Era Context handles the data pull automatically each time. No manual check required. ## Cross-agent memory: your context follows you Era Context shares memory across every connected AI client. Tell Poke something once — a savings goal, a budget preference, context about a transaction — and every other connected agent knows it too. Switch to Claude or ChatGPT later, and they already have your context. You can ask any agent to forget something, and it's gone everywhere. Your memory is private to you — never shared with other users, never used to train models. ## Is it safe to text your financial data? Your bank credentials never pass through Era or Poke — they're handled entirely by MX during the initial connection. After that, Era holds read access to your account data, protected by: - **AES-256 encryption** at rest - **TLS 1.3** in transit - **Revocable access** — disconnect Poke from Era Context at any time, instantly - **Activity log** — every action Era takes on your behalf is recorded and visible to you Era Financial Advisors LLC is SEC-registered (CRD #334404). For a deeper look at how the security model works, see [Is it safe to give AI your financial data?](/articles/is-it-safe-to-give-ai-your-financial-data). ## What tier do you need? The **Basic** tier (free) gives Poke read-only access to two accounts. That covers balances, transaction search, spending insights, and cash flow — enough to answer most questions you'd ever text. For automation — creating rules, tagging transactions, getting write access — you'll want the **Organize** tier at $9.99/month. See the full breakdown on the [pricing page](/pricing). ## Not just Poke Era Context works with any MCP-compatible client. If you use Claude, ChatGPT, OpenClaw, Cursor, or any other agent that supports MCP, the same bank connection works for all of them. Connect your bank once through Era, and every agent you trust can access it. For setup guides specific to other clients, see: - [Connecting Claude to your bank account](/articles/how-to-connect-claude-to-your-bank-account) - [Connecting ChatGPT to your bank account](/articles/how-to-connect-chatgpt-to-your-bank-account) - [Connect any AI agent to your bank](/articles/how-to-connect-ai-to-your-bank-account) The phone in your pocket is now connected to your financial life. Ask it. ### 25 things to ask your AI agent about your finances Era Context connects your bank accounts to any AI assistant through the Model Context Protocol (MCP). Once connected, your AI can do more than answer questions about your spending. It can create rules, organise your transactions, set goals, and monitor your accounts — all from a single conversation. These 25 prompts are organised into five categories. Each one works in Claude, ChatGPT, OpenClaw, or any MCP-compatible client connected to Era Context. Copy, paste, and see what happens. --- ## Spending analysis ### 1. "Compare my dining spending this month vs last month" Era Context pulls your dining transactions for both periods and shows you the difference — total spend, number of transactions, and the percentage change. Useful for spotting lifestyle creep before it becomes a habit. ### 2. "What are my top 5 spending categories this year?" Your AI ranks your categories by total spend and shows you where your money is actually going. Most people are surprised by at least one category on the list. ### 3. "Show me every transaction over $200 in the last 60 days" A quick way to surface large purchases you might have forgotten about. Your AI lists them with merchant names, dates, and amounts across all connected accounts. ### 4. "What does my spending forecast look like for the rest of this month?" Era Context analyses your spending patterns and recurring charges to project what you will likely spend between now and month-end. Helpful for deciding whether to hold off on a big purchase. ### 5. "Which day of the week do I spend the most money?" Your AI breaks down your spending by day of the week. Friday and Saturday tend to lead, but the specifics vary — and knowing your pattern is the first step to changing it. --- ## Savings and goals ### 6. "I want to save $5,000 by December — what do I need to set aside monthly?" Your AI calculates the monthly savings target based on how many months remain. Simple arithmetic, but having it in context with your actual cash flow makes it actionable. ### 7. "Remember that my emergency fund target is $10,000" This saves the goal to Era Context's cross-agent memory. Every AI client you use — Claude, ChatGPT, or any other — will know about this target going forward. When you ask about savings progress next month, the context is already there. ### 8. "What is my average monthly cash flow over the last six months?" Your AI calculates total income minus total spending for each of the last six months and gives you the average. This is the number that tells you how much room you actually have to save. ### 9. "Remember that I get paid on the 1st and 15th of each month" Another cross-agent memory prompt. Once your AI knows your pay schedule, it can give better advice about timing purchases and setting aside savings. ### 10. "Are there any months in the last year where I spent more than I earned?" Your AI checks each month's cash flow and flags the ones where outflows exceeded inflows. Knowing which months are tight helps you plan ahead for the next time. --- ## Rules and automation ### 11. "Create a rule that tags all transactions over $100 as 'big purchase'" Your AI builds the rule and presents it for your approval. Once active, every future transaction over $100 gets automatically tagged. You can then ask "show me all my big purchases this quarter" and get an instant answer. ### 12. "Set up a rule to alert me if any subscription increases in price" Your AI creates a monitoring rule that watches for price changes in your recurring charges. You approve it before it activates, and it runs automatically from that point on. ### 13. "Create a rule that tags all grocery transactions across all my accounts" Useful when grocery purchases are spread across multiple cards. The rule identifies grocery merchants and applies a consistent tag regardless of which account you used. ### 14. "Show me the rules I have active right now" Your AI lists every active rule with the original plain-English description you used to create it. Each rule carries an audit trail — you can see exactly what you asked for and when. ### 15. "Create a rule that flags any transaction from a merchant I have never used before" A practical security and awareness measure. New merchants get flagged so you can confirm they are legitimate. Approve the rule, and it runs quietly in the background. --- ## Housekeeping ### 16. "Clean up my merchant names — group Uber Eats, UberEats, and UBER EATS together" Bank feeds are messy. The same merchant appears under different names depending on the card, the terminal, or the phase of the moon. Your AI normalises them so your spending data is actually useful. ### 17. "Find all recurring charges and list them by amount" A quick audit of everything you are paying for on a regular basis, sorted from most to least expensive. The total at the bottom is usually the number that gets your attention. ### 18. "Tag all my coffee shop transactions this year" Your AI searches for coffee-related merchants across all accounts and tags them. You can then track your coffee spend as its own category — which is either reassuring or alarming. ### 19. "What categories are my transactions sorted into right now?" A snapshot of your current category structure. Useful for identifying gaps — transactions that are uncategorised or mis-categorised — before setting up rules to fix them. ### 20. "Find any duplicate or near-duplicate transactions in the last 30 days" Sometimes the same charge appears twice due to merchant processing. Your AI scans for transactions with the same amount and similar dates from the same merchant. Worth checking monthly. --- ## Strategic ### 21. "Which of my credit cards has the lowest utilisation?" Your AI checks your credit card balances against their limits and tells you which card has the most available headroom. Useful when you are deciding which card to use for a large purchase. ### 22. "What would happen to my monthly cash flow if I cancelled my three most expensive subscriptions?" Your AI identifies the three costliest recurring charges, sums them, and shows you the impact on your monthly bottom line. Sometimes the answer is enough to make you pick up the phone and cancel. ### 23. "Compare my total spending this year versus last year, broken down by quarter" A high-level view of whether your spending is trending up, down, or staying flat. Your AI pulls the numbers for each quarter and calculates the year-over-year change. ### 24. "What percentage of my income goes to fixed costs versus discretionary spending?" Your AI separates recurring charges and essentials from discretionary spending and calculates the split. Financial advisors call this your "fixed cost ratio" — and most people have no idea what theirs is. ### 25. "Based on my spending patterns, what are three realistic areas where I could cut back?" Your AI analyses your transaction history and suggests specific categories where your spending is higher than usual or where small changes would have a meaningful impact. Not generic advice — recommendations based on your actual data. --- ## How to get started These prompts work with any MCP-compatible AI client connected to Era Context. The setup takes less than five minutes: 1. Sign up at era.app and connect your bank accounts through MX. 2. Connect your AI client to Era Context at `https://context.era.app`. 3. Start asking questions. The Basic tier is free and supports 2 accounts with read-only access — enough to try any of the analysis prompts above. The Organize tier ($14.99/month) adds the rules engine, full read-write access, and support for up to 15 accounts. The Automate tier ($29.99/month) adds money transfers and rule-triggered automation. Every interaction is logged in your activity trail. Your data is encrypted with AES-256 at rest and TLS 1.3 in transit. Bank credentials are never stored. Your data is never sold and never used for advertising. Copy a prompt, paste it into your AI client, and see what your financial data can do when your AI can actually reach it. ### How financial advisors can use AI with client data Your AI assistant can access a client's financial data — with their explicit consent, scoped to specific accounts, and with every interaction logged. Era Context makes this possible through the Model Context Protocol (MCP), giving advisors a structured, auditable way to bring AI into their practice. Era Financial Advisors LLC is SEC-registered (CRD #334404). This article describes how the technology works, not legal or compliance advice. Consult your compliance team before adopting any new tooling. ## The problem with advisor prep today Before a client meeting, most advisors spend 15 to 30 minutes pulling up account summaries, scanning recent transactions, and trying to spot anything worth discussing. You log into a custodian portal, export a CSV, open a spreadsheet, and do the same thing you did last quarter. AI should be able to do this for you. The barrier has always been access — how do you give an AI assistant structured access to a client's financial data without handing over credentials or violating privacy expectations? ## How shared views work Era Context lets any user create a shared view — a read-only window into specific accounts. A client who uses Era can grant their advisor access to selected accounts. The advisor sees only what the client has chosen to share. Nothing more. Shared views are: - **Read-only.** The advisor can query data but cannot modify transactions, create rules, or move money. - **Scoped.** The client selects which accounts to include. A client with five accounts can share two. - **Revocable.** The client can remove access at any time, from any device. - **Audited.** Every query the advisor's AI agent makes against the shared view is logged in the activity trail. This is not a workaround or a screen-share. It is a first-class feature designed for exactly this use case. ## What an advisor can ask their AI Once a shared view is connected, the advisor can use any MCP-compatible AI client — Claude, ChatGPT, OpenClaw, or any other — to query the client's data in natural language. Here are some practical examples. ### Pre-meeting prep - "Summarise this client's financial activity for the last 90 days" - "What are the largest transactions this quarter?" - "Show me their recurring charges sorted by amount" - "Has their spending in any category increased by more than 20% compared to last quarter?" ### Cash flow review - "What is their average monthly cash flow over the last six months?" - "Are there any months where outflows exceeded inflows?" - "What does their spending forecast look like for the next 30 days?" ### Subscription and recurring charge audit - "List all recurring charges across their shared accounts" - "Have any subscriptions changed price recently?" - "What is the total monthly cost of all recurring charges?" ### Spending patterns - "What are their top five spending categories this year?" - "Compare their dining and entertainment spending this month versus last month" - "Flag any transactions over $500 in the last 60 days" Each of these queries runs through Era Context's 33 MCP tools. The AI client sends the request, Era Context processes it against the shared view, and the response comes back — all within the conversation. ## The audit trail Every interaction between an AI agent and Era Context is logged. The activity log records what was asked, when it was asked, and which accounts were queried. This is not a summary or a digest. It is a line-by-line record of every tool call. For advisors operating under regulatory oversight, this matters. You have a record of exactly what data your AI accessed and when. The client has the same visibility. ## Cross-agent memory for ongoing relationships Era Context includes cross-agent memory. If an advisor tells Claude "this client's goal is to pay off their car loan by March," that context persists. The next time the advisor opens ChatGPT and asks about the same client, that goal is already known. This eliminates the repetitive context-setting that makes AI assistants frustrating for ongoing relationships. Your AI remembers the client's goals, preferences, and financial context — across conversations and across AI clients. Memory is private to the advisor's account. It is never shared with other users and never used to train models. ## Setting it up The setup is straightforward: 1. The client connects their bank accounts to Era Context through MX, which supports thousands of financial institutions. 2. The client creates a shared view and selects which accounts to include. 3. The client shares the view with their advisor. 4. The advisor connects their AI client to Era Context at `https://context.era.app`. 5. The advisor can now query the shared view in natural language. The advisor does not need access to the client's login. The advisor does not see accounts the client has not shared. The client stays in control. ## What this is not This is not a replacement for your custodian, your financial planning software, or your compliance system. Era Context gives your AI assistant structured access to client financial data. It is one tool in your practice — the tool that makes your AI actually useful for the work you do every day. It is also not a robo-advisor. There is no automated investment advice, no portfolio rebalancing, no trade execution through shared views. This is data access for human advisors who want to use AI to work faster. ## Pricing for advisor use cases Era Context offers four tiers. For advisor workflows involving shared views: - **Basic** (free) supports 2 accounts and 1 view — suitable for evaluating the product. - **Organize** ($14.99/month) supports 15 accounts and 5 views with full read-write MCP access. - **Automate** ($29.99/month) adds money transfers and rule-triggered automation with 2 shared views. - **Optimize** ($49.99/month) includes unlimited shared views — designed for advisors managing multiple client relationships. Each client would maintain their own Era account and choose their own tier based on the features they need. ## Getting started If you are an RIA or independent advisor exploring how AI can fit into your practice, Era Context is worth evaluating. Connect your own accounts first, try the prompts above, and see what your AI can do with structured financial data. When you are ready to bring clients in, shared views make it possible without compromising privacy or auditability. Era works with any MCP-compatible client — Claude, ChatGPT, OpenClaw, and dozens more. You are not locked into a single AI provider, and neither are your clients. ### Era vs. Monarch vs. Copilot vs. YNAB: 2026 comparison Era, Monarch Money, Copilot, and YNAB take fundamentally different approaches to helping you manage your money. This comparison covers what each does well, where each falls short, and who each is best for — with a focus on AI integration, since that's where the landscape has shifted most dramatically. One caveat before we start: competitor features and pricing change. Everything listed here about Monarch, Copilot, and YNAB is based on their public websites and documentation at time of writing. If something looks wrong, it probably changed after this was published. ## The quick version | Feature | Era | Monarch Money | Copilot | YNAB | |---|---|---|---|---| | AI integration | MCP-native — works with any AI client (Claude, ChatGPT, Gemini, etc.) | Built-in AI assistant | Built-in AI assistant | No AI features at time of writing | | Your choice of AI | Yes — any MCP-compatible client | No — their assistant only | No — their assistant only | N/A | | Cross-agent memory | Yes — tell one AI, every AI knows | N/A | N/A | N/A | | Automation rules | Plain-English rules, AI-created, human-approved | Rule-based automations | Some automation | Rule-based categorization | | Shared/household views | Yes (paid tiers) | Yes | Limited at time of writing | Yes (multi-user) | | Budgeting methodology | AI-driven insights, no rigid envelope system | Flexible budgeting | Spending insights | Zero-based budgeting (envelope method) | | Free tier | Yes — Basic (2 accounts, read-only MCP) | No free tier at time of writing | No free tier at time of writing | Free trial only | | Paid pricing | $14.99–$49.99/mo | ~$14.99/mo at time of writing | ~$11.99/mo at time of writing | ~$14.99/mo at time of writing | | Platform | MCP server + web | Web + mobile apps | iOS + Mac (Apple only at time of writing) | Web + mobile apps | ## Monarch Money Monarch is a well-built, comprehensive personal finance platform. It does a lot of things right. **What Monarch does well**: Monarch offers clean dashboards for net worth tracking, budgeting, and investment monitoring. It supports account aggregation across a wide range of institutions. It has a collaborative mode for households, and its budgeting tools are flexible enough to work whether you prefer envelope-style or category-based tracking. At time of writing, Monarch also offers a built-in AI assistant for asking questions about your finances. **Where it differs from Era**: Monarch's AI is a built-in feature — an assistant inside the Monarch app. You use the model Monarch chose, inside Monarch's interface. If you prefer Claude over whatever model Monarch uses, or if you want to ask about your finances from within VS Code or another tool you use throughout your day, that's not how Monarch is designed to work. Era takes the opposite approach. There is no built-in chatbot. Your financial data is exposed through MCP, so you bring your own AI. Claude, ChatGPT, Gemini, OpenClaw — whatever you already use. Your money meets you where you are, not inside a dedicated finance app. **Best for**: People who want a traditional, well-polished finance dashboard with AI as one feature among many. ## Copilot Copilot (the finance app, not GitHub Copilot) has built a loyal following, particularly among Apple users. **What Copilot does well**: Copilot's design is excellent. It's one of the best-looking finance apps available, with thoughtful data visualizations and a clean iOS experience. It handles transaction categorization and spending tracking well, and at time of writing, it includes AI-powered insights within the app. **Where it differs from Era**: At time of writing, Copilot is primarily an Apple ecosystem product — iOS and Mac, with limited or no support for Android or Windows. Its AI features, like Monarch's, live inside the app. You can't connect Copilot's data to an external AI tool. Era is platform-agnostic by design. Because MCP is a protocol, not an app, Era works anywhere MCP is supported — which at this point includes AI clients on every major operating system. You're not choosing a finance app; you're adding a financial data layer to whatever tools you already use. **Best for**: Apple users who want a beautifully designed, self-contained finance app with integrated AI. ## YNAB (You Need a Budget) YNAB is the elder statesman of personal finance software, with a devoted community and a specific philosophy about money. **What YNAB does well**: YNAB's zero-based budgeting methodology genuinely changes how people think about money. Every dollar gets a job. The approach has helped millions of people get out of debt and build savings. YNAB's community, educational resources, and workshops are unmatched. It supports account syncing and manual entry, and its multi-platform support is solid. **Where it differs from Era**: YNAB is built around a specific budgeting methodology. If zero-based budgeting clicks for you, YNAB is exceptional at it. If it doesn't, the app can feel rigid. At time of writing, YNAB does not offer AI features — it's a manual, intentional approach to money management. Era is less opinionated about methodology. It doesn't enforce envelope budgeting or any other framework. Instead, it gives your AI access to your financial data and lets you interact with it however you want. You might ask your AI to build a zero-based budget, or you might ask it to just flag unusual spending. The AI adapts to your style. **Best for**: People who want a structured, proven budgeting methodology and value hands-on engagement with their finances. ## The real differentiator: MCP-native architecture The comparison table above captures features, but the deeper difference is architectural. Monarch, Copilot, and YNAB are apps. They have screens, dashboards, and workflows. Some have added AI as a feature inside those apps. Era is an MCP server. It's built from the ground up as a data layer for AI, not as an app with AI bolted on. This distinction matters in practice: **Any AI, not just theirs.** When you use Monarch's AI, you use Monarch's AI. When you use Era, you use whatever AI you want. Claude today, ChatGPT tomorrow, the next breakthrough model next week. One connection URL (`https://context.era.app`), every client. **AI that knows your context.** Era's cross-agent memory means your financial context travels with you. Tell Claude you're saving for a house. Switch to ChatGPT. It already knows. This isn't possible with a built-in chatbot — by definition, that chatbot's memory stays inside that app. **AI where you already are.** You don't open Era to ask about your money. You ask about your money wherever you already are. In the middle of a conversation with Claude about travel planning: "What can I actually afford for this trip?" In VS Code while reviewing a freelance contract: "What's my average monthly income from freelance work?" Your finances are context, not a destination. **Plain-English automation.** Era's rules engine lets you describe automation in natural language to any connected AI: "Categorize all DoorDash transactions as dining out." Your AI creates the rule, you approve it before it activates. A full audit trail records the original words you used. ## Cross-agent memory: a category of one This deserves its own section because no competitor offers anything like it at time of writing. When you use Monarch's AI assistant and mention that you're saving for a house, that information lives inside Monarch. If you switch to a different tool — or even a different conversation — you start over. Era's cross-agent memory works differently. Tell Claude you're saving $2,000 a month for a down payment. Open ChatGPT the next day and ask about your savings progress. It already knows. Open Gemini a week later and ask whether you can afford a vacation without derailing your savings plan. It already knows, too. Your financial context is stored in Era Context, not in any individual AI client, so it follows you everywhere. You can also ask any connected AI to forget something, and it's removed across all clients instantly. Your memory is private — never shared with other users, never used to train models. This matters because people don't use just one AI. You might prefer Claude for complex analysis, ChatGPT for quick questions, and Cursor for financial work while coding. With Era, you don't sacrifice context when you switch. With a built-in chatbot, you do. ## What Era doesn't do (yet) Honesty cuts both ways. Here's what Era doesn't offer that competitors do: - **Native mobile app**: Era is accessed through your AI client and the web. Monarch, Copilot, and YNAB all have dedicated mobile apps. - **Manual transaction entry**: YNAB in particular is built for people who want to manually enter transactions as a mindfulness practice. Era is built for automation. - **Budgeting templates**: Era doesn't have pre-built budget categories or envelope systems. Your AI can create these for you, but there's no guided setup wizard. - **Investment tracking dashboards**: Monarch offers detailed investment portfolio views. Era provides investment data through MCP tools, but the visualization happens in whatever AI client you're using. ## Security comparison All four platforms take security seriously, but the approaches differ. Era uses AES-256 encryption at rest and TLS 1.3 in transit. Bank credentials are never stored — they're handled entirely by MX, a regulated financial data provider, during authentication. Every AI agent interaction requires explicit authorization, and you can revoke access to any client at any time. A full activity log shows everything any AI has done on your behalf. Era is SEC-registered (Era Financial Advisors LLC, CRD #334404). Monarch, Copilot, and YNAB each use bank-level encryption and regulated aggregators for account connections. Check their current security pages for specifics, as implementations evolve. The key difference with MCP-based access is granularity: because each MCP tool has defined inputs and outputs, your AI can only perform specific, documented actions — not browse raw data or access internal systems. And because you control which clients connect, you can audit and revoke access per client rather than all-or-nothing. ## Pricing Era offers a free Basic tier with two connected accounts and read-only MCP access. Paid tiers start at $14.99/month (Organize) and go up to $49.99/month (Optimize), adding features like unlimited accounts, a full rules engine, read-write MCP access, shared views, and more. Monarch, Copilot, and YNAB pricing varies — check their current websites for the latest. At time of writing, all three charge a monthly or annual subscription with no permanent free tier. The free tier matters. You can connect two accounts to Era, wire it up to Claude or ChatGPT, and experience what MCP-native finance feels like before spending anything. If you've never asked an AI a question about your own financial data and gotten a real answer back, the free tier is the fastest way to understand why this matters. ## Who should choose Era Era is the right choice if: - You already use AI tools throughout your day and want your finances to be part of that workflow. - You have opinions about which AI model you use, and you don't want a finance app choosing for you. - You value automation over manual tracking. - You want your financial context to persist across conversations and across AI clients. - You want a free tier to start with. Era is probably not the right choice if: - You want a traditional finance dashboard you open and browse. - You prefer manual, intentional transaction entry as a budgeting practice. - You don't use AI tools regularly (yet). The landscape is shifting fast. The question is less "which finance app has the best AI feature?" and more "do you want your finance app to choose your AI, or do you want to choose your own?" If the answer is the latter, Era is currently the only option built from the ground up to support that. ### The future of personal finance is agent-native Every personal finance app ever built has asked the same thing of you: come here. Open this app. Look at this dashboard. Tap these buttons. The entire category — from Mint to Monarch to YNAB — is built on the assumption that managing money means going somewhere specific to do it. But you're already talking to AI all day. You ask Claude to help plan a trip. You ask ChatGPT to review a contract. You ask Gemini to summarize your week. What if your money just showed up in those conversations when it was relevant? Not because you opened a finance app, but because your AI already knows your financial picture. That's agent-native finance. And it changes everything about how you interact with your money. {{connect-claude-button}} ## The app era is ending Think about how you actually manage money today. You open your banking app to check a balance. You open a budgeting app to see if you're on track. You open a spreadsheet to plan a big purchase. You open your email to check a bill. Four apps, four contexts, four places where a sliver of your financial life lives. Now think about how you use AI. You have a conversation. You ask a question. You get an answer. You follow up. The conversation has context — the AI remembers what you said three messages ago, connects it to what you're asking now, and reasons about the whole picture. The gap between these two experiences is the opportunity. Finance apps give you data. AI gives you understanding. The problem is that AI can't understand your finances if it can't see them. ## What agent-native means Agent-native is not "an app with a chatbot." That's what most finance platforms are building right now — take the existing app, add a chat window, pipe questions to an LLM that can query the database. It works, sort of. But it's fundamentally limited. An app with a chatbot is still an app. You still have to go there. You still use their model. The chatbot can't connect your finances with your calendar, your emails, your documents, your code, or any other context your AI has access to. Agent-native means building the financial intelligence layer first, and letting the interface be whatever AI the person already uses. Three components make this work: **Context is the layer.** Era Context is a personal MCP server — a structured, secure data layer that exposes your financial accounts, transactions, insights, and memory to any AI that supports the Model Context Protocol. It's not an app you open. It's a layer your AI connects to. **Agency acts on your behalf.** Beyond reading data, agent-native finance means your AI can take action. Describe a rule in plain English — "tag every Uber charge as transportation" — and your AI creates it. You approve before it activates. Every rule remembers the exact words you used to create it. You're always in control, but you're not doing the tedious work. **Your AI is the interface.** There's no Era chatbot. Claude is your interface. Or ChatGPT. Or Gemini. Or OpenClaw. Or Cursor. Or whatever you use next month. One connection URL, every client. Your AI of choice already knows how to have a conversation, reason about data, and help you make decisions. Era gives it the financial data to reason about. ## What this looks like in practice Abstract architecture is less interesting than concrete moments. Here are some. **Morning coffee.** You're chatting with Claude about your day. You ask, "What's my financial situation looking like this week?" Claude checks your accounts, sees that rent cleared yesterday, notes that your credit card autopay is coming up on Friday, and tells you what's left. You didn't open an app. You asked a question in a conversation you were already having. **Planning a purchase.** You're looking at flights for a trip. You ask your AI, "Can I afford this?" Your AI checks your account balances, looks at your upcoming bills, considers the savings goal you mentioned last month (which it remembers, because Era's cross-agent memory persists across conversations and clients), and gives you an honest answer. Not a generic budgeting tip — a specific answer based on your specific numbers. **Spotting a problem.** Your AI notices a charge you haven't seen before. Or it notices you're being charged by a streaming service you told it you cancelled. It doesn't wait for you to open a dashboard and scroll through transactions. It mentions it when you're talking. "By the way, you're still being charged $15.99/month by that streaming service you said you cancelled in March." **Building a system.** You tell your AI, "Create a rule that categorizes all grocery store transactions and tags anything over $200 as worth reviewing." Your AI talks to Era, creates the rule, and shows it to you for approval. From now on, it just happens. You described what you wanted in plain English, and your financial system adapted. **Switching AIs.** You've been using Claude, but you want to try ChatGPT for a while. You add one line of configuration. ChatGPT connects to Era Context, and it already knows everything — your accounts, your goals, your preferences, your rules. You told Claude you're saving for a down payment. ChatGPT knows it too. No re-onboarding, no re-explaining, no starting over. ## Why now Three things converged to make agent-native finance possible in 2026. **MCP reached critical mass.** The Model Context Protocol went from a niche standard to a widely-supported protocol. Claude, ChatGPT, Gemini, VS Code, Cursor, and dozens of other clients now support MCP. This means building an MCP server is no longer a bet — it's a viable distribution strategy. Build once, connect to every AI client. **AI got good enough.** Large language models can now reason about financial data accurately and helpfully. They can spot patterns, compare periods, forecast trends, and explain findings in plain language. Two years ago, you wouldn't trust an AI to analyze your bank transactions. Today, it's one of the things AI does best. **People are already living in AI.** The average knowledge worker has multiple conversations with AI per day. AI is already the place where thinking happens — planning, analyzing, deciding. Finance is one of the last categories that hasn't met people there. ## What this is not Agent-native finance is not a suggestion that you should blindly trust AI with your money. Every automation rule in Era requires your explicit approval before it activates. Every AI agent interaction requires explicit authorization. You can revoke access to any client at any time. There's a full activity log showing everything any AI agent has done on your behalf. It's also not a replacement for financial advisors, accountants, or professional guidance. Era is SEC-registered (Era Financial Advisors LLC, CRD #334404), and takes regulatory obligations seriously. Agent-native finance is a better interface for your financial data — not a substitute for professional judgment when you need it. ## The three products Era is building three products around the agent-native thesis. **Era Context** is live today. It's your personal MCP server — 33 tools across seven groups that give your AI access to your accounts, transactions, insights, memory, and automation. The Basic tier is free, with two connected accounts and read-only MCP access. **Era Agency** is a companion platform for AI-driven financial automation at a deeper level. It's on the waitlist, not shipped yet. **Era Thesis** is an AI algorithmic trading platform. Also on the waitlist. Context is the foundation. It's the layer that makes everything else possible. And it's available now. ## Try it If you've read this far and the idea resonates, the fastest way to understand it is to experience it. Sign up at [era.app](https://era.app), connect a bank account, and {{connect-claude}} — then ask your AI a question about your money. The first time your AI answers with real data — your actual balance, your actual spending, your actual recurring charges — the shift clicks. Finance stops being a place you go and starts being a thing your AI knows. That's agent-native. And once you've felt it, the old way feels like checking your email by driving to the post office. ### Subscription creep: how AI detects the charges you forgot about You are paying for things you forgot you signed up for. A streaming service you tried for a month. A productivity app you used twice. A gym membership at a gym you moved away from. The average person underestimates their recurring charges by 40 to 50 percent. The charges are small enough to ignore individually but large enough to matter collectively. Era Context connects your bank accounts to any AI assistant through the Model Context Protocol (MCP). One prompt is all it takes to surface every recurring charge across all your connected accounts. ## The "show me all my subscriptions" moment Connect your accounts to Era Context, open your preferred AI client — Claude, ChatGPT, OpenClaw, or any MCP-compatible tool — and type: > "Show me all my recurring charges" Your AI scans every connected account and returns a complete list of recurring charges. Not just the ones you remember. All of them. Grouped by category, sorted by amount, with the total at the bottom. This is the moment most people say "I'm spending how much on subscriptions?" The number is almost always higher than expected. ## What the AI surfaces Era Context identifies recurring charges across your connected accounts. When you ask, your AI can: - **List every recurring charge** with the merchant name, amount, and frequency. - **Group charges by category** — streaming, software, fitness, food delivery, news, and so on. - **Calculate your total monthly subscription spend** across all accounts. - **Flag charges that changed price** since they first appeared. - **Identify charges you may have forgotten** — services that bill annually or quarterly are easy to lose track of. - **Spot duplicate charges** — two Spotify accounts, an old and new Netflix plan running simultaneously, overlapping cloud storage subscriptions. This is not a static report. It is a conversation. You can ask follow-up questions: "When did that Hulu charge start?" or "How much have I paid to Adobe in total this year?" ## How recurring charge detection works Era Context connects to your bank accounts through MX, which supports thousands of financial institutions. Once connected, Era Context's 33 MCP tools give your AI structured access to your transaction data. Recurring charge detection works by analysing your transaction history across all connected accounts. The AI identifies patterns — charges that repeat at regular intervals from the same merchant. This catches monthly subscriptions, annual renewals, quarterly fees, and irregular-but-recurring charges like insurance premiums. Because Era Context has access to all your connected accounts, it catches charges you might miss if you only check one bank statement. That gym membership on your old debit card. The app subscription billing to a credit card you rarely use. ## Setting up a subscription watchdog Finding your current recurring charges is useful. Catching new ones as they appear is better. Era Context includes a rules engine that you can set up in plain English. Ask your AI: > "Create a rule that alerts me whenever a new recurring charge appears" Your AI creates the rule and presents it for your approval. Nothing activates without your say-so. Once approved, the rule monitors your transactions and flags any new recurring pattern it detects. You can also create rules for specific scenarios: - "Alert me if any subscription increases in price" - "Tag all recurring charges under $10 as 'micro-subscriptions'" - "Flag any recurring charge I haven't used the associated service for" Each rule remembers the exact words you used to create it, giving you a clear audit trail of what you asked for and when. ## The subscription audit: a step-by-step walkthrough Here is a practical workflow for a complete subscription audit. Open your AI client with Era Context connected and work through these prompts: **Step 1: Get the full picture** > "List all my recurring charges across all accounts, grouped by category, with monthly totals" **Step 2: Find the surprises** > "Which of these recurring charges have I been paying for more than a year?" > "Are any of these charges duplicates or overlapping services?" **Step 3: Quantify the impact** > "What is my total monthly subscription spend? What would it be if I cancelled the three most expensive ones?" **Step 4: Check for price increases** > "Have any of my subscriptions increased in price in the last six months?" **Step 5: Set up ongoing monitoring** > "Create a rule that alerts me whenever a new recurring charge appears on any account" > "Create a rule that alerts me if any existing subscription increases in price" This entire process takes about five minutes. Without AI, it would take an afternoon of logging into individual bank accounts, downloading statements, and cross-referencing charges in a spreadsheet. ## Cross-agent memory keeps your context When you identify subscriptions you want to keep track of, you can tell your AI to remember that context: > "Remember that I want to keep Netflix, Spotify, and my gym membership, but I should review everything else quarterly" This is stored in Era Context's cross-agent memory. The next time you ask about subscriptions — in Claude, ChatGPT, or any other connected client — your AI already knows which charges are intentional and which ones need scrutiny. ## What this costs Era Context's Basic tier is free and supports 2 connected accounts with read-only access. For a full subscription audit across all your accounts, the Organize tier at $14.99 per month supports 15 accounts with full rules engine access and unlimited categories and tags. If the audit saves you even one forgotten $15 subscription, the product pays for itself in the first month. ## Getting started 1. Sign up at era.app and connect your bank accounts through MX. 2. Connect your preferred AI client to Era Context at `https://context.era.app`. 3. Ask: "Show me all my recurring charges." 4. Review the list. Cancel what you do not need. 5. Set up a rule to catch new charges going forward. You are almost certainly paying for something you forgot about. Now you have a way to find it. ### Your agent, your choice: why Era is client-agnostic The AI landscape moves fast. Twelve months ago, most people used one AI assistant. Today, many use three or four. New clients launch constantly — Claude, ChatGPT, OpenClaw, Manus, Gemini, Cursor, Perplexity — each with different strengths, different interfaces, different personalities. Now imagine locking your entire financial life into one of them. That's exactly what most AI-powered finance tools do. They build a chatbot inside their app, connect it to one model, and call it "AI-powered." Your data lives inside their interface. Your conversations live inside their model. Switch to a different AI and you lose everything. Era doesn't work that way. ## The Model Context Protocol changes everything Era Context is built on the Model Context Protocol (MCP) — an open standard that lets AI clients connect to external data sources securely. Instead of building a chatbot, Era built a personal MCP server for your finances that any compatible client can connect to. What this means in practice: your bank accounts, your transaction history, your spending insights, your rules, your financial memory — all of it is accessible from whatever AI client you prefer. Claude today. ChatGPT tomorrow. OpenClaw next week. Your data doesn't move. Your AI does. This isn't a philosophical position. It's architecture. Era Context is the server. Your AI client is the client. Swap the client anytime. The server — your financial data, your rules, your memory — stays exactly where it is. ## Why client lock-in is a bad deal Consider what happens when a traditional fintech app builds AI into their product. They pick a model — usually whatever's cheapest or whatever partnership they can strike. They wrap it in their interface. And now you're stuck with their choices. **Their model, not yours.** Maybe they chose a model that's great at summarizing but terrible at nuanced financial planning. You have no say. **Their interface, not yours.** Maybe you prefer Claude's conversation style, or ChatGPT's data visualization, or Cursor's developer-friendly interface. Too bad — you get their chat window. **Their update cycle, not yours.** When a better model launches, you wait for them to integrate it. If they don't, you're out of luck. **Their survival, not yours.** If the company pivots, gets acquired, or shuts down, your financial AI history goes with it. Era eliminates all of this. Your financial data lives in Era Context. You bring whatever AI client you want to the conversation. If a better client launches tomorrow, connect it. If your current favorite stumbles, switch. Your data, rules, and memory stay with Era regardless. ## Cross-agent memory is the proof Client-agnostic architecture sounds abstract until you experience cross-agent memory. Tell Claude your savings goal. Open ChatGPT. It already knows. Tell ChatGPT you want to track a new spending category. Open Gemini. It's already there. This only works because Era Context is the single source of truth. Your financial memory doesn't live in Claude's conversation history or ChatGPT's session storage. It lives in Era. Every connected agent reads from and writes to the same context. Ask any agent to forget something and it's gone everywhere. Not just from that agent's memory — from your entire financial context, across every client. This is what client-agnostic really means. Not just "we have multiple integrations." But your entire financial life working identically across every AI you use, with shared memory, shared rules, and shared context. ## The walled-garden alternative Compare this with what most competitors offer. Traditional fintech AI follows a familiar playbook: download our app, use our chatbot, stay inside our walls. Your conversations are trapped in their interface. Your insights are generated by their model. Your automations only work inside their product. Some offer a decent experience. But every one of them is a bet — a bet that this particular company, with this particular model, inside this particular interface, will be the best option for your finances not just today, but for years to come. That's a bad bet. The AI space is evolving too fast for any single provider to stay on top indefinitely. The model that's best at financial reasoning today might be second-best in six months. The interface you love now might feel dated by next year. Era doesn't ask you to make that bet. Use the best AI available right now. When something better comes along, switch. Your financial life doesn't skip a beat. ## Which clients work with Era Era Context works with any MCP-compatible client. The list grows constantly, but here are popular ones people use today: **Claude** — Anthropic's assistant, available as Claude Desktop and Claude Code. Strong at nuanced reasoning and detailed financial analysis. **ChatGPT** — OpenAI's assistant. Widely used, strong at conversational interaction and data visualization. **OpenClaw** — An open-source MCP client that's gaining rapid adoption. Worth trying if you prefer open-source tools. **Gemini** — Google's assistant. Strong integration with the Google ecosystem. **Manus** — An autonomous AI agent designed for complex, multi-step tasks. **Cursor, VS Code, GitHub Copilot** — Developer-focused clients that work well for people who live in their code editor and want to manage finances without switching contexts. **Perplexity** — A research-focused AI, useful for financial research and comparison tasks. This isn't an exhaustive list. Any client that supports MCP can connect to Era Context. New clients launch regularly, and every one of them works automatically — no integration work needed from Era, no waiting for updates, no compatibility concerns. ## What stays with Era When you switch between AI clients, here's what travels with you: **Your connected accounts.** Every bank account, credit card, and financial institution you've linked stays connected. No re-linking, no re-authenticating. **Your transaction history.** Every categorization, every tag, every merchant rename you've set up. Nothing resets. **Your rules.** Every automation rule you've created — whether you built it in Claude, ChatGPT, or any other agent — continues to run. Rules live in Era Context, not in any client. **Your financial memory.** Every goal, preference, plan, and note you've shared with any agent. Your context is continuous across every client. **Your activity log.** A complete record of everything Era Context has done on your behalf, regardless of which agent initiated it. The only thing that changes is the AI you're talking to. Everything else is persistent, portable, and yours. ## Security across every client A client-agnostic model raises a fair question: if any AI can access my financial data, how is that secure? Every AI agent interaction requires your explicit authorization. You control exactly which clients have access to your Era Context. Connect Claude — Claude gets access. Don't connect Gemini — Gemini sees nothing. You can revoke access to any client at any time, instantly. If you stop using a particular AI, disconnect it. It loses access to your financial data immediately. Your data is encrypted with AES-256 at rest and TLS 1.3 in transit. Bank credentials are never stored by Era. Your data is never sold, never used for advertising, and never shared without your explicit permission. The security model is simple: Era holds your data. You decide which AIs can see it. You can change your mind at any time. ## The future is multi-agent We're heading toward a world where people use different AI clients for different tasks. Maybe Claude for deep financial planning. ChatGPT for quick spending checks. A specialized agent for tax optimization. Another for investment research. In that world, locking your financial data inside one client makes as little sense as storing all your files on one computer with no cloud backup. Your financial context needs to be portable, persistent, and independent of any single AI. That's what Era Context provides. A stable, secure, personal financial server that works with whatever AI landscape evolves around it. The agents will change. The models will improve. New clients will launch and old ones will fade. Your financial data, your rules, and your memory stay with Era — and follow you wherever you go. Era Context is available now. Connect it to your preferred MCP-compatible client and bring your entire financial life with you — to any AI, at any time. ### The complete guide to AI-powered financial automations You already tell your AI assistant what you want. "Summarize this document." "Draft a reply." Now imagine saying: "Tag every coffee shop transaction as 'daily habits'" — and having it actually happen across your bank accounts, automatically, from now on. That is what Era's automation rules do. You describe what you want in plain English. Your AI agent creates the rule. You approve it. Era handles the rest. ## The old way versus the new way Most people manage their finances reactively. You open your banking app, scroll through a wall of transactions, and mentally sort them: that was groceries, that was a subscription, that was a reimbursement. Some people build spreadsheets. Some tag things in budgeting apps. All of it is manual work you repeat every week. Era flips this. Instead of you doing the sorting, you tell your AI what to sort, and it builds a rule that runs on every transaction going forward. The rule remembers the exact words you used to create it, so you always have an audit trail of your intent. Here is what that looks like in practice: - "Flag any subscription over $30" — your AI creates a rule that watches for recurring charges above that threshold and tags them for your review. - "Tag all coffee shops as 'daily habits'" — every transaction at a coffee merchant gets categorized automatically. - "Clean up merchant names so I can actually read them" — messy bank descriptions like "SQ *BLUEBOTTLE COF" become "Blue Bottle Coffee." - "Detect any recurring charge I haven't seen before" — new subscriptions get flagged before they become invisible line items. You are not learning a rule builder. You are not configuring filters and dropdowns. You are having a conversation. ## How rules get created The process is deliberately simple, because the point is that you should not need to think about the mechanics. **Step one: describe what you want.** Open any AI assistant connected to Era Context — Claude, ChatGPT, OpenClaw, or any other MCP-compatible client. Tell it what you want to automate. Use natural language. Be as specific or as vague as you like. "I want to track how much I spend eating out each month" works. So does "categorize everything from DoorDash, Uber Eats, and Grubhub as 'delivery food.'" **Step two: your AI creates the rule.** Based on your description, your AI agent constructs a rule within Era Context. The rule captures your intent — the categories to watch, the conditions to match, the actions to take. **Step three: you approve it.** Nothing activates without your explicit approval. Your AI presents the rule for review, and you confirm before it goes live. This is non-negotiable. Era never acts on your finances without your say-so. **Step four: the rule runs.** From that point forward, every matching transaction gets processed automatically. New transactions are evaluated as they arrive. You can check the activity log at any time to see exactly what each rule has done. ## The pre-built rule library Not everyone wants to start from scratch. Era includes a library of pre-built rules you can browse and activate with a single tap. These cover common use cases that most people need: - **Merchant name cleanup** — transform cryptic bank descriptions into readable names. - **Category detection** — automatically sort transactions into spending categories like dining, groceries, transportation, and entertainment. - **Income identification** — recognize payroll deposits and other income sources so your cash flow view separates money in from money out. - **Transfer detection** — distinguish transfers between your own accounts from actual spending, so moving money from checking to savings does not inflate your expense totals. - **Recurring charge identification** — detect subscriptions and recurring bills based on transaction patterns. Each pre-built rule is isolated to your account. You can customize it, deactivate it, or replace it with your own version at any time. ## Custom rules through conversation The pre-built library covers the basics. Custom rules are where things get personal. Because you create rules through conversation with your AI, you can express nuance that no dropdown menu could capture. Some examples of rules people create: **Spending awareness rules:** - "Flag any single transaction over $200 that is not rent or a bill payment." - "Tag transactions at gas stations as 'commute costs' on weekdays and 'travel' on weekends." - "Detect when my dining spending in a single week exceeds $150." **Organization rules:** - "Create a tag called 'side hustle expenses' and apply it to any transaction at Adobe, Canva, or my web hosting provider." - "Separate my wife's gym membership from mine — hers is Planet Fitness, mine is Equinox." - "Tag anything from Amazon under $20 as 'impulse buys.'" **Detection rules:** - "Alert me to any new recurring charge I have not explicitly approved." - "Flag duplicate charges — same merchant, same amount, same day." - "Detect when a subscription price increases from what it was last month." You do not need to know the "right" way to phrase these. Your AI interprets your intent and builds the appropriate rule. If something is ambiguous, it asks you to clarify. ## The approval step: nothing happens without your say-so This is worth emphasizing because it matters for trust. Every rule — whether created by your AI from a conversation or activated from the library — requires your explicit approval before it takes effect. When your AI proposes a rule, you see exactly what it will do. You can modify it, reject it, or approve it. Once approved, the rule runs automatically. But the initial activation is always a conscious decision you make. You can also deactivate any rule at any time. There is no lock-in, no "are you sure?" friction. Rules are tools that work for you, and you stay in control of which ones are active. ## What rules can do today Rules in Era are focused on transaction organization and awareness. Here is a concrete list of what they handle: - **Auto-categorize transactions** by merchant, amount, pattern, or description. - **Tag transactions** with custom labels you define. - **Detect recurring charges** and surface new subscriptions. - **Flag anomalies** — unusual amounts, unexpected merchants, potential duplicates. - **Clean up merchant names** so your transaction history is readable. - **Identify income** and separate it from expenses in your cash flow view. - **Detect transfers** between your own accounts to avoid double-counting spending. Rules operate across all your connected accounts. A rule that tags coffee shops will catch your credit card latte and your debit card cold brew alike. ## Cross-agent rule creation One of the more interesting aspects of Era's architecture is that rules are not tied to the AI agent that created them. If you create a rule through Claude, it applies to your Era Context account. Later, if you are chatting with ChatGPT, that agent can see the rule, modify it, or create new ones. This works because Era Context is the system of record. Your AI agents are interfaces to it, not silos. Tell Claude to "tag all Costco transactions as 'bulk shopping.'" Next week, ask ChatGPT to "show me all my active rules." It will see the Costco rule you created through Claude. Your rules, your memory, your financial context — all of it persists across agents and across conversations. Every conversation picks up where the last one left off, regardless of which AI you are talking to. ## The activity log: full transparency Every action a rule takes is recorded in your activity log. You can see: - Which rule fired and when. - What transaction it acted on. - What action it took (categorized, tagged, flagged). This is not buried in a settings page. The activity log is a first-class feature of Era Context, showing everything that happens with your money — syncs, rule executions, balance changes, and more. If you ever wonder "why is this transaction tagged this way?" the answer is one tap away. ## Rules that work across your whole financial picture A single rule applies across every connected account. This matters more than it sounds. Most budgeting tools are account-centric. You set up categories for your credit card, then do it again for your debit card, then again for your second credit card. Rules in Era are account-agnostic. "Tag all dining transactions" catches the restaurant on your Visa, the food truck on your debit, and the business dinner on your Amex. One rule, total coverage. This also means rules catch things that single-account views miss. A rule that detects recurring charges will surface subscriptions regardless of which card they bill to. If Netflix charges your Visa and Spotify charges your checking account, both show up. You get a complete picture without manually cross-referencing accounts. ## When rules get smarter over time Rules in Era are not static filters. The automation engine has been continuously improved — better pattern matching, more accurate category detection, smarter income identification, and better handling of edge cases like recurring charges with unusual merchant names. Recent improvements include better detection of transfers between your own accounts (so moving money to savings does not look like spending), more accurate identification of bounced payments, and cross-account pattern detection that spots related transactions across multiple accounts. You do not need to update your rules when these improvements ship. The engine gets smarter; your existing rules benefit automatically. ## Building a system, not just setting rules The real power of automations is not any single rule. It is the system you build over time. Start simple. Activate a few rules from the library — merchant name cleanup and category detection are good first picks. Let them run for a week. Check your transactions and see how they look. Then start adding your own. Notice that you keep manually categorizing Lyft rides? Create a rule. Annoyed that your bank calls your electric bill "UTILPAY ACH DEBIT"? Clean it up. Want to track how much you spend on your dog? Tag pet stores, vet visits, and Chewy orders. Over weeks, your transaction history transforms from a wall of inscrutable text into a clear, organized picture of where your money goes. And you did not build a spreadsheet or learn a new app. You just had a few conversations with your AI. ## Getting started To use automation rules, you need an Era account with the Organize plan ($14.99/month) or higher. The free Basic plan includes five template rules from the library. Organize unlocks the full rules engine and unlimited custom rules. Connect your bank accounts through any MCP-compatible AI agent, or directly at [era.app](https://era.app). The MCP connection URL is `https://context.era.app` — add it to Claude, ChatGPT, OpenClaw, or whichever client you use. Once connected, just tell your AI what you want to automate. It handles the rest. You approve. The rules run. Your finances organize themselves. You have been doing manually what your AI could handle. Now you know. ### Pattern-based rules any agent can create for you You're chatting with your AI about your spending. You scroll through some transactions and notice a pattern — you hit the same three coffee shops every week. So you say: "I want to tag all my coffee shop visits as 'daily habits'." Your agent says: "Done — I've created a rule for that. Want me to apply it to your existing transactions too?" That's it. No forms. No settings pages. No learning a rules UI. You described what you wanted, and it happened. ## Rules through conversation, not configuration Every budgeting app has a rules system buried somewhere in its settings. You navigate through menus, fill out forms, pick from dropdowns, configure conditions. It works, technically. But nobody would call it enjoyable. Era Context takes a different approach. Your rules engine is a conversation. "Tag all transactions from Uber and Lyft as 'transportation'." Done. "Categorize anything from Whole Foods, Trader Joe's, or Sprouts as groceries." Done. "Flag any transaction over $200 that I haven't tagged yet." Done. You talk to your AI the way you'd talk to a person who manages your books. Describe the pattern. Describe what should happen. Your agent creates the rule and waits for your approval. ## Nothing activates until you say so This is the part that matters most. Your AI agent creates rules based on your instructions, but nothing activates until you explicitly approve it. When your agent creates a rule, you see exactly what it will do. Which transactions it will match. What action it will take. The rule sits in a pending state until you say "yes, activate that" or "actually, let's adjust it." This approval step isn't a speed bump — it's the point. You're delegating the tedious work of setting up rules to your AI while keeping full control over what actually runs against your financial data. The agent does the configuration work. You make the decisions. ## What rules can do Rules in Era Context cover the repetitive housekeeping that makes financial data actually useful. **Auto-categorize transactions.** "Put all my subscription services into a 'subscriptions' category." Your agent identifies the pattern, creates the rule, and every matching transaction gets categorized automatically — past and future. **Tag spending patterns.** "Tag everything at restaurants during the weekend as 'social spending'." Tags give you a flexible layer on top of categories. Use them however you want — by mood, by purpose, by project, by person. **Clean up merchant names.** Those cryptic transaction descriptions from your bank — "SQ *JOES COFFEE #1247" — can be cleaned up automatically. "Whenever you see a transaction from SQ *JOES COFFEE, rename the merchant to Joe's Coffee." Readable statements, zero effort. **Detect recurring charges.** "Show me any new recurring charge that appears for the first time." Stop subscription creep before it starts. Your rules can flag new recurring patterns so you're always aware of what's billing you. **Flag anomalies.** "Alert me if my spending in any category jumps more than 50% compared to last month." Let your rules watch for unusual patterns while you focus on everything else. ## The pre-built library Not everyone wants to start from scratch. Era Context includes a library of pre-built rules you can browse and activate with a single conversation. Ask your agent: "Show me what rules are available." You'll see templates for common patterns — subscription detection, merchant cleanup, category organization, spending alerts. Pick the ones that make sense for your situation, customize them if you want, and activate them. The library is a starting point, not a constraint. Every pre-built rule can be modified, and you can always create completely custom rules by describing what you want. ## Every rule remembers your words Here's a detail that sounds small but matters more than you'd expect: every rule stores the exact words you used to create it. Six months from now, you might look at a rule and wonder why it exists. Instead of trying to reverse-engineer the logic from technical conditions, you'll see your original request: "I created this because I wanted to track how much I spend on my dog." This is an audit trail that actually makes sense. Not timestamps and technical logs, but your own words explaining your own intent. When your AI helps you review your rules later, it can reference why you created each one, not just what each one does. ## Any agent can create rules Because Era Context works with any MCP-compatible client, you can create rules from whichever AI you're using. Start a rule in Claude. Modify it in ChatGPT. Review it in Gemini. Your rules live in Era Context, not in any single agent. This also means you get the benefit of each AI's conversational strengths. If one agent is better at understanding nuanced descriptions, use that one for complex rules. If another is faster for quick changes, use that for maintenance. Your rules don't care which agent created them. ## The shift from managing to describing The bigger picture here is about how you interact with your financial tools. Traditional apps ask you to learn their interface — their forms, their menus, their vocabulary. Era Context asks you to use your own words. "I want to separate business meals from personal dining" is a complete instruction. Your agent figures out which merchants qualify, creates the rule, and waits for your go-ahead. You described an outcome. Your agent handled the implementation. This is what financial automation should feel like. Not clicking through configuration screens. Not learning a rules syntax. Just saying what you want and approving the result. ## Getting started with rules Rules are available on Era Context's Organize tier and above. Connect your accounts, start talking to your AI, and describe what you want organized. Your agent handles the rest — you just approve. Start simple. "Tag all my grocery stores." See how it feels. Then get specific. "Flag any subscription I'm paying for that I haven't used in 30 days." Then get creative. The only limit is what you can describe. Your financial data should work the way you think about it, not the way a database structures it. Rules are how you bridge that gap — one conversation at a time. ### Using Era Context in a developer workflow "How much did I spend on AWS this month?" You can ask that question without leaving your terminal. If you use Claude Code, Cursor, VS Code with GitHub Copilot, or any other MCP-compatible development tool, your coding assistant already speaks the same protocol as Era Context. One configuration line, and your AI has secure access to your financial data alongside your codebase. ## Why this matters for developers Developers live in their tools. Switching to a banking app to check a balance or review transactions is a context switch — small, but frequent. If you are already talking to an AI assistant while you code, that same assistant can answer financial questions without you opening a new tab. This is not about building fintech. It is about convenience. You are mid-flow, you wonder if that AWS bill posted yet, and you ask. Your AI checks Era Context, gives you the answer, and you keep coding. Same conversation, no app switching. The connection works because Era Context is a Model Context Protocol (MCP) server. MCP is an open standard for giving AI assistants access to external data sources. Your dev tools already support it — Claude Code, Cursor, and VS Code use MCP to connect to codebases, databases, and APIs. Era Context is just another MCP server, except it connects to your bank accounts instead of your database. ## Setup: Claude Code Claude Code supports MCP servers natively. Add Era Context to your configuration: 1. Open your Claude Code MCP settings. 2. Add a new MCP server with the URL: `https://context.era.app` 3. Authenticate with your Era account when prompted. That is it. Claude Code now has access to your financial data. Ask it anything: "What is my checking account balance?" "Show me my recurring subscriptions." "How much did I spend on food this week?" Claude Code will use Era Context's tools automatically when your question is about finances and your codebase when your question is about code. No mode switching required. ## Setup: Cursor Cursor supports MCP servers through its settings panel: 1. Open Cursor settings and navigate to the MCP section. 2. Add a new server with the URL: `https://context.era.app` 3. Authenticate with your Era account. Once connected, you can ask financial questions inline while coding. Cursor routes the query to Era Context and returns the answer in the same chat panel you use for code assistance. ## Setup: VS Code with GitHub Copilot VS Code supports MCP through GitHub Copilot: 1. Open VS Code settings and find the MCP server configuration. 2. Add Era Context with the URL: `https://context.era.app` 3. Authenticate when prompted. GitHub Copilot will route financial queries to Era Context and code queries to your workspace — seamlessly, in the same conversation. ## Setup: any MCP-compatible client The pattern is the same for every MCP-compatible tool: add `https://context.era.app` as an MCP server and authenticate. Era works with any client that supports the Model Context Protocol — the list includes Gemini, Perplexity, Cline, OpenClaw, and more. If your tool can connect to an MCP server, it can connect to Era Context. ## What you can ask Once connected, your coding AI has access to 33 tools across 7 groups within Era Context. You do not need to know the tool names — just ask questions in natural language, and your AI selects the right tools automatically. **Account queries:** - "What is my checking account balance right now?" - "List all my connected accounts and their balances." **Transaction searches:** - "How much did I spend on AWS this month?" - "Show me all transactions over $100 this week." - "What recurring subscriptions am I paying for?" **Spending analysis:** - "Compare my spending this month versus last month." - "Break down my spending by category for March." - "What is my daily average spend this week?" **Financial context:** - "What are my financial goals?" (if you have told any AI agent about them — Era remembers across agents). - "What did I tell Claude about my savings plan?" (cross-agent memory means any connected agent can recall this). **Rule management:** - "Create a rule to tag all GitHub and Vercel charges as 'dev tools.'" - "Show me my active automation rules." ## Cross-agent memory in practice Here is where it gets interesting for developers who use multiple AI tools. Era Context has cross-agent memory: tell one AI something about your finances, and every other connected AI knows it too. Say you are using Claude on your phone and mention, "I am saving for a new MacBook Pro — target is $3,000 by September." That fact is stored in Era Context. Later, you are in Cursor working on a project and ask, "How am I tracking toward my savings goals?" Cursor queries Era Context, retrieves the MacBook Pro goal you told Claude about, and gives you an update — all without you repeating yourself. This works because Era Context is the memory layer, not any individual AI. Your preferences, goals, and financial context persist across every connected client. You can also ask any agent to forget something, and it is gone everywhere. ## Privacy and access control A reasonable question: "Should I connect my bank accounts to my dev tools?" Here is how the security works. Your dev tool connects to Era Context, not to your bank. Era Context connects to your bank through MX, a regulated financial data aggregator. Your bank credentials are never stored by Era — the connection uses OAuth, so you authenticate directly with your bank. Your AI agent sees structured financial data — balances, transactions, categories. It does not see raw database rows, full account numbers, or bank credentials. Every interaction requires explicit authorization. You can revoke access to any client at any time. Data is encrypted with AES-256 at rest and TLS 1.3 in transit. Your financial data is never sold, never used for advertising, and never used to train AI models. You choose what to share and with whom. If you want Cursor to have access but not VS Code, you control that. Each client connection is independent. ## A practical developer day Here is what a typical day might look like with Era Context connected to your dev tools: **9:00 AM** — You start coding in Cursor. Mid-morning you wonder if your freelance client's payment hit yet. "Did I receive any deposits over $2,000 this week?" Cursor checks Era Context: yes, the payment posted yesterday. **12:30 PM** — Over lunch in Claude Code, you ask: "What is my total spend on SaaS subscriptions this month?" Claude lists your recurring charges — Vercel, GitHub, Notion, Linear, Figma. You notice you are still paying for a tool you stopped using. You ask Claude to flag it. **4:00 PM** — You switch to ChatGPT on your phone for a non-coding question. You ask: "How much have I spent on dev tools this year?" ChatGPT pulls the same data, sees the tag you created through Cursor, and gives you the total. You did not re-explain anything. That is the workflow. Your financial data is there when you need it, invisible when you do not, and consistent across every tool you use. ## Getting started Create an Era account at [era.app](https://era.app). The free Basic plan gives you read-only MCP access with two connected accounts — enough to try the workflow. The Organize plan ($14.99/month) adds full read-write MCP access, unlimited categories and tags, and the full rules engine. Add `https://context.era.app` to your dev tool's MCP configuration. Authenticate. Start asking questions. Your coding AI already knows your code. Now it knows your money too. ### Introducing cross-agent memory for your finances You told Claude about your savings goal last Tuesday. You mentioned a budget target for dining out. You noted that your partner handles the mortgage and you cover utilities. Small details, scattered across conversations, building up a picture of your financial life. Today you open ChatGPT. You ask it to review your spending for the month. And it already knows about the savings goal. It already knows about the dining budget. It already knows who pays what. Wait — it remembered? That's cross-agent memory. And it changes everything about how personal finance works with AI. ## What cross-agent memory actually means Most AI conversations are isolated. You tell Claude something, and that knowledge lives in Claude. You tell ChatGPT something, and that lives in ChatGPT. Switch between them and you're starting from scratch every time, re-explaining your situation, repeating your preferences, restating your goals. Era Context sits between you and every AI agent you use. When you mention something about your finances — a goal, a preference, a plan, a note to yourself — Era remembers it. Not in one agent's conversation history. In your Era Context, which every connected agent can access. Tell Claude you're saving for a down payment. Open ChatGPT tomorrow and ask for a spending review. It will factor in your savings goal without you saying a word. Open Gemini next week and ask whether you can afford a vacation. It already has the context. One memory. Every agent. Always current. ## The moments that matter Cross-agent memory sounds like a technical feature. It isn't. It's about the small human moments where AI stops feeling like a tool and starts feeling like it actually knows you. **The savings goal that follows you.** You're chatting with Claude about your finances and casually mention you want to save $20,000 for a home down payment by next March. Two days later, you ask ChatGPT to analyze your spending. It doesn't just show you numbers — it frames everything around your savings goal, highlighting where you could cut back to hit your target. **The preference that sticks.** You tell your AI that you consider your gym membership essential, not discretionary. From that point on, every agent respects that classification. No more arguing with your budgeting tool about whether $50 a month for fitness is a luxury. **The plan that evolves.** You start a financial plan in one conversation and refine it in another. Maybe you're working through debt paydown with Claude, then switch to ChatGPT to explore investment options once you're debt-free. Each conversation builds on the last because your context is continuous. **The note to your future self.** You can tell any agent to remember something for later. "Remind me that I negotiated a lower rate on my car insurance — it renews in September." Months later, any agent can surface that note when September arrives. ## How it works for you There's no setup. No configuration page. No import/export step. When you're talking to any AI agent connected to Era Context, you just talk. Say what matters to you. Your agent stores it in your Era Context automatically. "I want to keep my dining spending under $400 this month." Remembered. "My partner and I split rent 60/40." Remembered. "I'm thinking about switching jobs — want to build up three months of expenses first." Remembered. Every agent you connect to Era Context can access these memories. Your financial context is always complete, always current, no matter which AI you happen to be using. ## Forgetting is just as important Memory without control isn't a feature — it's a liability. You can ask any agent to forget something, and it's gone everywhere. Not just from that conversation. From your entire Era Context. Every connected agent loses access to that piece of information instantly. "Forget that I was considering switching jobs." Done. No trace across any agent. This matters because your financial life changes. Goals shift. Plans evolve. Relationships change. Your financial context should change with them, and you should be the one deciding what stays and what goes. ## Privacy by design Your memory is yours. Period. Your financial context is private to you. It is never shared with other users. It is never used to train AI models. Era encrypts your data with AES-256 at rest and TLS 1.3 in transit. Bank credentials are never stored. Every AI agent interaction requires your explicit authorization. You can revoke access to any client at any time. If you disconnect Claude, Claude loses access to your memory. If you disconnect ChatGPT, same thing. Your data stays with Era, not with any single AI provider. ## Why this matters now The AI landscape is moving fast. New clients launch every month. The best AI for financial conversations today might not be the best one next year. Claude, ChatGPT, OpenClaw, Manus, Gemini — the list keeps growing. Without cross-agent memory, every time you try a new AI client, you start over. You lose context. You lose the relationship you've built with your financial data. With Era Context, switching is painless. Your goals, preferences, plans, and notes follow you. Try a new AI client and it already knows your financial life because your memory lives in Era, not in any single agent. This is what "personal" in personal finance should actually mean. Not personal to one app. Not personal to one AI. Personal to you. ## Getting started Cross-agent memory is live in Era Context today. Connect your Era Context to any MCP-compatible client — Claude, ChatGPT, OpenClaw, Gemini, and dozens more — and start talking about your finances. If you're ready to set everything up right now, [how to set up cross-agent memory for your finances](/articles/how-to-set-up-cross-agent-memory) walks through it step by step — from creating your account to building your first memory profile. Once you're up and running, [what Era Context stores and how to manage it](/articles/what-era-context-stores-and-how-to-manage-it) covers exactly what gets saved and how to view, edit, or remove anything in your profile. The first time your second agent remembers something you told your first agent, you'll feel it. That small moment of surprise: "Wait, it remembered?" That's the moment personal finance becomes truly personal. ### Era's privacy model: your data, your control "Is it safe to connect my bank to AI?" Yes. But you should not take our word for it. Here is exactly how Era handles your financial data — what we store, what we never store, what your AI can see, and how you stay in control at every step. ## What Era connects to and how Era connects to your bank through MX, a regulated financial data aggregator that works with thousands of financial institutions. When you link a bank account, you authenticate directly with your bank through MX's secure connection flow. Era never sees, handles, or stores your bank credentials — username, password, security questions, or multi-factor codes. Those stay between you and your bank. The connection uses OAuth where your bank supports it. You grant MX permission to access specific account data, and MX provides that data to Era. You can revoke this permission at any time through Era or directly through your bank. ## What your AI can see When you connect an AI agent to Era Context, it does not get raw access to a database. It sees structured financial summaries: account balances, transaction lists with merchant names and categories, spending breakdowns, recurring charges, and the financial context you have shared (goals, preferences, notes). Here is what your AI can access: - **Account balances** — current balances for your connected accounts. - **Transaction data** — merchant name, amount, date, category, and any tags you have applied. - **Spending analysis** — breakdowns by category, time period comparisons, cash flow summaries. - **Your financial context** — goals, preferences, and facts you have told any connected AI agent (stored in Era's cross-agent memory). - **Automation rules** — the rules you have created and their activity. - **Activity log** — a record of everything Era Context has done on your behalf. Your AI accesses this data through Era Context's 33 MCP tools. Every tool call is a discrete, authorized action — not a bulk data dump. ## What never leaves Era Some data never reaches your AI agent at all: - **Bank credentials** — never stored by Era, never accessible to any agent. - **Full account numbers** — your AI sees masked account identifiers, not full account or routing numbers. - **Raw database rows** — your AI receives structured summaries, not unfiltered data exports. - **Other users' data** — your financial data is isolated to your account. No agent, no employee, and no other user can access it. ## OAuth consent: you choose what to share Every AI agent that connects to Era Context goes through an authorization step. You explicitly grant each client permission to access your financial data. This is not a blanket "connect everything" switch — you authorize each client independently. Want Claude to have access but not ChatGPT? That is your choice. Want to give Cursor read-only access for quick balance checks but full access to Claude for rule creation? You control the scope. Each connection is independent. Authorizing one client does not authorize any other. And revoking one does not affect the rest. ## Revoke any agent, any time If you change your mind about a connected AI agent, you revoke its access instantly. No waiting period, no "please contact support," no data retention after disconnection. Revoke access, and that agent can no longer query your financial data. This applies to every connected client — Claude, ChatGPT, Cursor, OpenClaw, or any other MCP-compatible tool. You stay in control of which agents have access at all times. ## The activity log: full audit trail Era Context maintains a complete activity log of every action taken on your behalf. Every time an AI agent queries your balance, searches your transactions, creates a rule, or modifies a tag — it is recorded. The activity log shows: - **What happened** — the specific action taken. - **When it happened** — timestamp for every event. - **Which agent did it** — so you know if it was Claude, ChatGPT, or another client. This is not a hidden system log. It is a feature you can access at any time within Era Context. If you ever wonder what an AI agent did with your data, the answer is there. ## What Era never does with your data Some commitments are absolute: - **Your data is never sold.** Not to advertisers, not to data brokers, not to anyone. - **Your data is never used for advertising.** Era does not serve ads, and your financial data is never used to target ads on any platform. - **Your data never trains AI models.** Your transactions, balances, and financial context are not used as training data for any machine learning model. - **Your data is never shared without your explicit permission.** No third party sees your financial data unless you have specifically authorized a connection. ## Encryption and infrastructure The technical specifics matter for people who care about them: - **AES-256 encryption at rest** — your data is encrypted when stored. - **TLS 1.3 in transit** — your data is encrypted when moving between your device, Era, and your bank. - **Era Financial Advisors LLC** is SEC-registered (CRD #334404). This is a regulated entity with compliance obligations, not a side project. - **Brokerage services** (when applicable) are provided through Alpaca Securities LLC, a FINRA/SIPC member. ## Cross-agent memory and privacy Era's cross-agent memory lets you tell one AI agent something and have every connected agent know it. "I am saving $500 a month for a house down payment" — tell Claude, and ChatGPT knows it too. This memory is private to you. It is never shared with other users, never used to train models, and never accessible to anyone but your authorized agents. You can ask any agent to forget something, and it is removed everywhere — from Era's memory and from every connected agent's context. You control what your agents remember. You control what they forget. The memory exists to make your experience better, not to build a profile for someone else. ## Addressing the real concern The question "is it safe to connect my bank to AI?" is really two questions. **"Is my bank connection secure?"** Yes. Era does not store your bank credentials. The connection goes through MX, a regulated aggregator. You authenticate directly with your bank. You can revoke the connection at any time. **"Can I trust an AI with my financial data?"** That depends on the controls. With Era, every agent requires explicit authorization. Every action is logged. You can revoke access instantly. Your data is never sold, never used for ads, never used to train models. And you can see exactly what every agent has done in the activity log. Era does not ask you to trust blindly. It gives you the tools to verify. ## Getting started Create a free Era account at [era.app](https://era.app). The Basic plan includes read-only MCP access for up to two accounts — enough to evaluate the security model yourself before committing to a paid plan. Connect your AI agent using the MCP URL: `https://context.era.app`. Authorize it. Check the activity log. See exactly what happens. Your data, your control. That is the model. ### Connect any AI agent to your bank accounts Era Context connects your bank accounts, credit cards, and investments to any AI agent that supports the Model Context Protocol (MCP). One account. One setup. Every agent — Claude, ChatGPT, OpenClaw, Manus, Cursor, Gemini, and dozens more. You don't have to pick one AI and hope it's the right one. Era is agent-agnostic. Connect your bank once, and any MCP-compatible client can access your financial data with your explicit permission. ## How it works Era Context is a personal MCP server. MCP — the Model Context Protocol — is an open standard that lets AI agents connect to external data sources securely. Era implements this protocol for personal finance. The setup follows the same pattern regardless of which AI client you use: 1. **Create an Era account** at [era.app](https://era.app). The Basic tier is free. 2. **Connect your bank** through Era Context using MX, a regulated financial data provider. Your bank credentials are never stored by Era. 3. **Add the MCP server URL** to your AI client: `https://context.era.app` 4. **Complete the OAuth consent** — you explicitly authorize each AI agent to access your data. 5. **Ask your AI about your money** — it responds with real data. That's it. Five steps, five minutes, any agent. ## Popular AI clients Era works with any MCP-compatible client. Here are the popular ones and where to find their MCP settings: | Client | Where to add MCP servers | |--------|--------------------------| | Claude Desktop | Settings → MCP Servers | | Claude Code | MCP configuration file | | ChatGPT | Settings → MCP servers | | Cursor | Settings → MCP | | OpenClaw | MCP server configuration | | Manus | Agent settings → MCP | | Gemini | MCP server settings | | VS Code (Copilot) | MCP configuration | | Perplexity | MCP settings | | Cline | MCP configuration | | Hermes | MCP settings | For every client, the MCP server URL is the same: ``` https://context.era.app ``` No trailing slash. No path. No `/sse`. Just that URL. ### Configuration example Most clients accept a JSON config like this: ```json { "mcpServers": { "era-context": { "url": "https://context.era.app" } } } ``` Some clients have a visual settings panel where you paste the URL directly. Either way, the URL is `https://context.era.app`. ## What your AI can do once connected Era Context exposes 33 MCP tools across 7 groups. Your AI client picks the right tools automatically — you just ask questions in natural language: **Account basics:** - "What are my account balances?" - "Show me all my connected accounts" **Transaction search:** - "How much did I spend at Whole Foods this month?" - "List all transactions over $200 in the last 30 days" - "Find my recurring charges" **Spending analysis:** - "Break down my spending by category this month" - "Compare my spending this month to last month" - "What's my cash flow look like?" - "Forecast my spending for the rest of the month" **Financial context:** - "Can I afford a $2,000 purchase right now?" - "What's my daily financial summary?" These work in any connected agent. Ask Claude, ask ChatGPT, ask OpenClaw — the data is the same because it all flows through Era Context. ## Cross-agent memory This is the part that changes everything. When you tell one AI agent something about your finances — a savings goal, a spending preference, context about a transaction — Era Context remembers it. And that memory is available to every connected agent. Tell Claude you're saving for a down payment. Open ChatGPT later. It already knows. Switch to OpenClaw. Same context. Your financial memory follows you across agents. This isn't just convenient. It means you can use whichever AI is best for the task at hand without losing context. Use Claude for deep analysis, ChatGPT for quick questions, Cursor while coding — they all share the same understanding of your financial life. You control this memory completely: - Ask any agent to remember something, and every agent knows it - Ask any agent to forget something, and it's gone everywhere - Your memory is private to you — never shared with other users, never used to train models ## Frequently asked questions ### Is it safe to connect my bank account to an AI? Era doesn't give AI agents your bank credentials. Here's how the security works: - **Bank authentication** is handled entirely by MX, a regulated financial data provider. Your username and password are never stored by Era. - **Data encryption**: AES-256 at rest, TLS 1.3 in transit. - **Explicit authorization**: Every AI agent must go through OAuth consent before it can access anything. You approve each one individually. - **Data policy**: Your data is never sold, never used for advertising, never shared without your explicit permission. - **SEC-registered**: Era Financial Advisors LLC, CRD #334404. Built by a team from Stripe, Robinhood, CashApp, Apple, and Google. ### What data can the AI see? On the Basic (free) tier, your AI agents have read-only access to: - Account balances for up to two connected accounts - Transaction history - Basic financial metrics On paid tiers, additional capabilities include automation rules, full categorization, tags, and more accounts. The Organize tier ($14.99/month) adds 15 accounts, a full rules engine, and read-write MCP access. The Automate tier ($29.99/month) adds money transfers. The Optimize tier ($49.99/month) adds brokerage and investment capabilities. ### Can I revoke access? Yes, instantly. You can disconnect any AI agent from Era Context at any time. The agent immediately loses access to your financial data. You can also reconnect later without losing your transaction history or financial memory. ### Can the AI move my money? On the Basic and Organize tiers, access is read-only or read-write for organization purposes — no money movement. The Automate tier ($29.99/month) enables transfers, but every transfer rule requires your explicit approval before it activates. The AI can never move money without your permission. ### Do I need a separate Era account for each AI agent? No. One Era account works with every MCP-compatible client. Connect your bank once, authorize each agent via OAuth, and they all access the same data through your single Era Context. ### What happens if my bank connection drops? Era preserves your full transaction history even if a bank connection is temporarily disconnected. You can reconnect at any time without losing data. Era Context will notify you if a connection needs attention. ### Does Era work with my bank? Era supports thousands of financial institutions through MX. Most major banks, credit unions, credit card issuers, and investment platforms are supported. The best way to check is to start the bank connection flow in Era Context — you'll see whether your institution is available. ## Automation rules On the Organize tier and above, you can describe automation rules in plain English through any connected AI agent: - "Tag all Uber transactions as commuting" - "Flag any charge over $500" - "Clean up merchant names for all Amazon purchases" - "Detect recurring charges I might have forgotten about" Your AI agent creates the rule, and you review and approve it before it activates. Nothing runs without your sign-off. Every rule remembers the exact words you used to create it — a full audit trail of your intent. There's also a library of pre-built rules you can browse and activate directly from Era Context. ## Getting started 1. Sign up at [era.app](https://era.app) — the Basic tier is free 2. Connect your bank through Era Context 3. Add `https://context.era.app` to your AI client's MCP settings 4. Complete the OAuth consent 5. Ask your AI about your money One setup. Every agent. Your finances, your choice of AI. ### How to connect ChatGPT to your finances with Era Era gives ChatGPT direct access to your bank accounts, credit cards, and investments through the Model Context Protocol (MCP). Once connected, you can ask ChatGPT about your money the same way you ask it anything else — and it responds with real data, not generic advice. Setup takes about five minutes. Here's how. ## Why connect ChatGPT to your finances You already use ChatGPT to think through problems, draft plans, and make decisions. But when it comes to money, you've been on your own — copying numbers from bank apps, pasting spreadsheets, describing your situation from memory. Era Context changes that. It's a personal MCP server that sits between your financial accounts and your AI agents. Connect it once, and ChatGPT can see your balances, search your transactions, analyze your spending, and remember your financial goals. Your AI already knows how to reason about money. It just needs access. ## What you need - A ChatGPT account (Plus, Team, or Enterprise with MCP support) - A bank account at a supported institution (Era supports thousands through its banking data provider) - Five minutes ## Step 1: Create your Era account Go to [era.app](https://era.app) and sign up. The Basic tier is free — it includes two connected accounts with read-only MCP access, which is everything you need to get started. ## Step 2: Connect your bank From Era Context, start the bank connection flow. Era uses MX, a regulated financial data provider, to link your accounts securely. Your bank credentials are never stored by Era — MX handles them during the authentication process. Pick your bank, sign in, complete any multi-factor authentication, and you're connected. Most banks are live in under 30 seconds. ## Step 3: Add Era Context to ChatGPT In ChatGPT, open Settings and find the MCP server configuration. Add a new MCP server with this URL: ``` https://context.era.app ``` That's the full URL. No path, no trailing slash. ## Step 4: Authorize the connection ChatGPT will redirect you to an OAuth consent screen where you grant it permission to access your financial data through Era. Review the permissions, approve, and you're live. This authorization is explicit and revocable. You can disconnect ChatGPT from Era Context at any time. ## Step 5: Start talking to your money Open a new ChatGPT conversation and try one of these: **"What's my checking account balance right now?"** ChatGPT pulls the number from Era Context in real time. No plugins, no workarounds. **"How much did I spend on groceries this month?"** ChatGPT searches your transactions, totals the grocery category, and gives you the answer. **"Compare my spending this month to last month"** A side-by-side breakdown, pulled from your actual data. ## Conversations to try after connecting Here's where it gets interesting. These aren't hypothetical — they work right now with Era Context connected: **Budgeting check-in:** "I'm trying to keep my dining out under $300 this month. Where am I at?" **Subscription audit:** "List all my recurring charges. Are there any I'm paying for but probably not using?" **Cash flow planning:** "I get paid on the 15th and 30th. Based on my upcoming recurring charges, how much discretionary spending do I have until my next paycheck?" **Spending patterns:** "What are my top 5 spending categories over the last 3 months? How do they compare to the 3 months before that?" **Financial context:** "I'm thinking about whether I can afford a $1,200 vacation next month. Look at my income, recurring expenses, and current balances and tell me what you think." Each of these uses Era Context's MCP tools — 33 tools across 7 groups — to pull real data and give you real answers. ChatGPT selects the right tools automatically based on your question. ## Cross-agent memory Era Context includes cross-agent memory. When you tell ChatGPT something about your finances — "I'm saving for a down payment" or "I consider anything over $100 a big purchase" — Era remembers it. And that memory syncs across every AI agent you connect. Switch to Claude tomorrow, and it already knows your savings goal. No re-explaining. Your financial context follows you. You control this memory completely. Ask any agent to forget something, and it's gone everywhere. Your data is never shared with other users or used to train models. ## Troubleshooting **ChatGPT can't connect to the MCP server** Verify the URL is exactly `https://context.era.app`. No `/sse`, no extra path. **OAuth redirect doesn't work** Sign into your Era account at [era.app](https://era.app) in the same browser first, then retry. **Data seems stale or missing** New bank connections may take a moment to sync fully. Check Era Context directly — if your accounts and transactions appear there, they'll show up in ChatGPT shortly. **Limited functionality** The Basic (free) tier includes read-only access to two accounts. For automation rules, unlimited categories, and more accounts, check out the Organize tier ($14.99/month). ## Security Your financial data is protected with: - **AES-256 encryption** at rest - **TLS 1.3** in transit - **Bank credentials never stored** by Era - **Data never sold** or shared without your explicit permission - **Revocable access** — disconnect any AI agent anytime Era Financial Advisors LLC is SEC-registered (CRD #334404), built by a team from Stripe, Robinhood, CashApp, Apple, and Google. ## What's next You've connected ChatGPT to your bank accounts. Here's how to get more out of it: - **Connect additional accounts** — credit cards, investments, accounts at other banks - **Set financial goals** — tell ChatGPT what you're saving for and let cross-agent memory track it - **Try other agents** — Era works with any MCP-compatible client, including Claude, Cursor, OpenClaw, Manus, Gemini, and more. One setup, every agent. - **Explore automation** — on the Organize tier, you can create rules in plain English that auto-categorize transactions, flag unusual charges, and keep your finances tidy ChatGPT already knows how to think about money. Now it can see yours. ### Connecting Claude to your bank account with Era Era connects Claude to your bank accounts, credit cards, and investments through the Model Context Protocol (MCP), giving Claude secure, read-only access to your financial life in under five minutes. No CSV exports. No copy-pasting statements. Just ask Claude about your money, and it answers with real data. This guide covers setup for both Claude Desktop and Claude Code. By the end, you'll have Claude pulling live balances, analyzing spending patterns, and answering questions about your finances in natural language. {{connect-claude-button}} ## What you need before you start - A Claude Desktop or Claude Code subscription (any plan works) - A bank account, credit card, or investment account at a supported institution (Era supports thousands of financial institutions through its banking data provider) - About five minutes That's it. ## Step 1: Create your Era account Go to [era.app](https://era.app) and sign up. The Basic tier is free and includes two connected accounts with read-only MCP access — enough to follow this entire guide. You'll verify your email and land in Era Context, your personal financial hub. Era Context is where you manage connections, review activity, and control what your AI agents can access. ## Step 2: Connect your bank From Era Context, tap the connect flow to link a bank account. Era uses MX, a regulated financial data provider, to establish a secure connection to your institution. Your bank credentials are never stored by Era — they're handled entirely by MX during the authentication handshake. Select your bank, sign in with your online banking credentials, and complete any multi-factor authentication your bank requires. Most connections are live within 30 seconds. Once connected, you'll see your accounts and recent transactions appear in Era Context. If you have a checking and savings account at the same bank, both will show up automatically. ## Step 3: Add the MCP config to Claude Desktop Open Claude Desktop and navigate to Settings, then the MCP Servers section. You can also {{connect-claude}} directly from here. Add a new MCP server with the following configuration: ```json { "mcpServers": { "era-context": { "url": "https://context.era.app" } } } ``` That's the entire configuration. The URL is `https://context.era.app` — nothing else to add. Save your settings. Claude Desktop will attempt to connect to the Era Context MCP server. ## Step 4: Complete the OAuth consent When Claude Desktop connects to Era Context for the first time, you'll be redirected to an OAuth authorization screen. This is where you explicitly grant Claude permission to access your financial data through Era. Review the permissions, confirm, and you're connected. Every AI agent interaction requires this explicit authorization — nothing happens silently. ## Step 5: Try your first query Open a new conversation in Claude Desktop and try something like: **"What's my checking account balance?"** Claude will pull your live balance from Era Context and respond with the current number. No plugins to configure, no APIs to learn. Claude now has real financial context. Here are more things to try: - "How much did I spend on dining out this month?" - "Show me my recurring charges" - "Compare my spending this month to last month" - "What's my cash flow look like?" - "List all my transactions from Amazon this year" Each of these queries hits Era Context's MCP tools — 33 tools across 7 groups covering accounts, transactions, insights, automation, and more. You don't need to know the tools exist. Claude picks the right ones automatically based on your question. ## Setting up Claude Code If you use Claude Code (Anthropic's CLI for developers), the setup is just as fast. Add Era Context to your MCP configuration: ```json { "mcpServers": { "era-context": { "url": "https://context.era.app" } } } ``` Complete the OAuth flow when prompted, and you're live. Claude Code is particularly useful for developers who want to query financial data while working — checking whether a subscription charge went through, reviewing expenses before submitting a report, or building personal finance scripts with live data. Try asking Claude Code: - "Analyze my spending categories for the last 90 days" - "What recurring subscriptions am I paying for?" - "Show me my daily financial summary" ## Cross-agent memory: tell Claude once, every agent knows One of Era Context's most powerful features is cross-agent memory. When you tell Claude something about your finances — a savings goal, a budget preference, context about a transaction — Era Context remembers it. And that memory persists across every connected AI agent. Tell Claude "I'm saving for a house and trying to keep dining under $400/month." Switch to ChatGPT tomorrow, and it already knows. No re-explaining. Your financial context follows you across agents. You can also ask any agent to forget something, and it's gone everywhere. Your memory is private to you — never shared with other users, never used to train models. ## Troubleshooting **Claude says it can't find the MCP server** Double-check that the URL in your config is exactly `https://context.era.app`. No trailing slash, no path, no `/sse`. **OAuth screen doesn't appear** Make sure you're signed into your Era account in the same browser that handles the OAuth redirect. Try signing in at [era.app](https://era.app) first, then retry the connection in Claude. **Transactions aren't showing up** New bank connections can take a minute to fully sync. If your accounts appear in Era Context but transactions don't show in Claude, wait a moment and try again. If the bank connection shows as disconnected in Era Context, reconnect it — Era preserves your transaction history even when reconnecting. **"Permission denied" or similar errors** Your Era plan determines which MCP capabilities are available. The Basic (free) tier provides read-only access to two accounts. If you're hitting limits, check your plan in Era Context. **Multiple accounts at the same bank** Era pulls all accounts from a connected institution automatically. If you only want Claude to see specific accounts, you can choose which accounts are active in Era Context. ## What Claude can do with your financial data With Era Context connected, Claude can: - **Check balances** across all your connected accounts - **Search transactions** by merchant, category, amount, or date range - **Analyze spending** patterns and trends over any time period - **Compare periods** — this month vs. last month, this quarter vs. last quarter - **Forecast spending** based on your historical patterns - **Identify recurring charges** and subscriptions - **Provide cash flow summaries** showing money in vs. money out - **Remember your preferences** and financial goals across conversations On the Organize tier ($14.99/month), Claude can also help you create automation rules in plain English — "tag all Uber transactions as commuting" or "flag any charge over $500" — which you review and approve before they activate. ## Security and privacy Era takes security seriously: - **AES-256 encryption** at rest - **TLS 1.3** in transit - **Bank credentials are never stored** — they're handled by the regulated data provider - **Your data is never sold**, never used for advertising, never shared without your explicit permission - **Revoke access anytime** — disconnect any AI agent from Era Context whenever you want Era Financial Advisors LLC is SEC-registered (CRD #334404). This isn't a weekend project. It's a regulated financial service built by a team from Stripe, Robinhood, CashApp, Apple, and Google. ## Next steps You're connected. Claude can see your money. Here are some ways to go deeper: - **Connect more accounts** — add credit cards, investment accounts, or accounts at other banks - **Explore the Organize tier** — unlock automation rules, unlimited categories and tags, and full financial metrics - **Try other AI agents** — Era works with any MCP-compatible client, including ChatGPT, Cursor, OpenClaw, Manus, Gemini, and more. One Era account, every agent. - **Set financial goals** — tell Claude what you're working toward and let cross-agent memory keep every agent aligned The whole point of Era Context is that your AI already knows how to reason about money. It just needed access. Now it has it. ### What is the Model Context Protocol (MCP)? The Model Context Protocol (MCP) is an open standard that lets AI assistants — Claude, ChatGPT, Gemini, and others — connect directly to external data sources and tools. Instead of copying information into a chat window, MCP gives your AI secure, structured access to live data. Era is one of the first platforms to use MCP for personal finance, turning your bank accounts, credit cards, and investments into a data layer that any AI you already use can read and act on. If you've ever wished you could just ask your AI "what did I spend on groceries last month?" and get a real answer from real data, MCP is the technology that makes that possible. ## The problem MCP solves AI assistants are remarkably good at reasoning, planning, and explaining. But they have a fundamental limitation: they can only work with information you give them. Ask Claude about your finances, and it can offer generic budgeting advice. It can't tell you that your electricity bill jumped 40% this month, because it can't see your electricity bill. Historically, people solved this problem in clunky ways: - **Copy-pasting**: Export a CSV from your bank, paste it into the chat, hope the AI parses the columns correctly. - **Screen scraping**: Automated tools that log into your bank and pull data — fragile, often against terms of service, and a security nightmare. - **Proprietary chatbots**: Finance apps that build their own AI chatbot inside their app. You get AI, but only their AI, only inside their app, using only their model. Each approach has the same structural flaw: the data is trapped. Either it's trapped in a file format your AI has to guess at, trapped behind a scraper that breaks when the bank changes a button, or trapped inside an app that chose your AI for you. MCP eliminates the trap. ## How MCP works, simply MCP defines a standard way for AI clients (the apps you chat with) to connect to MCP servers (the systems that hold your data). Think of it like USB for AI — a universal connector that lets any compatible device talk to any compatible peripheral. An MCP server exposes **tools** — structured actions the AI can call. These are not free-text prompts. They're typed, documented endpoints: "list my bank accounts", "search transactions from last week", "analyze spending by category". The AI reads the tool descriptions, decides which ones to call based on your question, and presents the results in natural language. The key properties: - **Open standard**: Anyone can build an MCP server. Anyone can build an MCP client. No single company controls the protocol. - **Structured data**: The AI receives typed, clean data — not a blob of text it has to parse. This means better, more accurate answers. - **Permissioned access**: You decide which MCP servers your AI can connect to. You can revoke access at any time. - **Client-agnostic**: One MCP server works with every MCP-compatible client. Build once, connect everywhere. ## What this means for personal finance Personal finance is one of the best use cases for MCP. Your financial data is deeply personal, constantly changing, and incredibly useful for the kind of reasoning AI is good at — spotting patterns, forecasting, comparing, planning. Before MCP, getting AI to work with your money meant one of two things: 1. **Upload your data manually** every time you want an answer. Tedious, error-prone, and the AI forgets everything between sessions. 2. **Use a finance app's built-in chatbot**, which locks you into that app's AI model, that app's interface, and that app's idea of what questions you're allowed to ask. MCP creates a third option: your financial data becomes a persistent, secure layer that any AI you choose can access. You're not locked into any particular AI client. You're not uploading files. Your data stays with your data provider, and your AI queries it live. ## How Era uses MCP Era Context is a personal MCP server for your finances. You connect your bank accounts to Era through MX (a regulated financial data provider), and Era exposes that data as a set of 33 MCP tools across seven groups: accounts, transactions, insights, activity, billing, knowledge, and connections. Setting it up takes one line of configuration. In Claude Desktop, Cursor, VS Code, or any other MCP-compatible client, you point to `https://context.era.app` with your authentication token. That's it. Your AI can now see your money. Here's what becomes possible: - "What's my checking account balance?" — answered from live data, not memory. - "How does my restaurant spending this month compare to last month?" — computed from your actual transactions. - "Show me all recurring charges over $50" — pulled from pattern analysis across your accounts. - "Remember that I'm saving for a trip to Japan" — stored in Era's cross-agent memory, so every AI you connect knows about your goal. That last point is worth dwelling on. Era's cross-agent memory means that if you tell Claude about your savings goal, ChatGPT knows about it too. Your financial context persists across conversations and across AI clients. No re-explaining. ## Beyond querying: memory and automation MCP is not just about reading data. Because MCP tools can accept inputs as well as return outputs, an MCP server can offer write operations — actions your AI can take on your behalf. Era uses this to enable two capabilities that fundamentally change how you manage money: **Cross-agent memory.** Era Context includes a knowledge system that persists across conversations and across AI clients. Tell Claude that you're saving $500 a month for a trip to Japan. Later, open ChatGPT and ask "am I on track for my savings goal?" ChatGPT already knows about the trip, the target amount, and your progress — because Era stored that context, not any individual AI client. You can also ask any agent to forget something, and it's gone everywhere. Your memory is private to you, never shared with other users, and never used to train models. **Plain-English automation rules.** Describe a rule to any connected AI — "categorize all Starbucks transactions as coffee" or "tag any charge over $500 as worth reviewing" — and your AI creates it through Era's rules engine. You approve before it activates. Nothing runs without your explicit sign-off. Every rule remembers the exact words you used to create it, giving you a full audit trail in your own language. A library of pre-built rules is also available to browse and activate without writing anything from scratch. These capabilities are only possible because MCP provides a standard, structured way for AI to interact with external systems. Without MCP, each AI client would need its own custom integration — and your data would be siloed in whichever client you happened to use. ## MCP vs. the walled-garden chatbot Most finance apps that offer AI follow the same playbook: embed a chatbot inside the app. You open the app, tap the chat icon, and talk to whatever model they chose for you. This approach has real limitations: - **Model lock-in**: You use their model, not yours. If a better model launches tomorrow, you can't switch. - **Interface lock-in**: You have to be inside their app. You can't ask about your finances from Claude Desktop, or from VS Code while you're coding, or from whatever AI tool you use throughout your day. - **Context isolation**: The chatbot only knows what's inside that app. It can't connect your financial data with your calendar, your email, your project management tools, or anything else your AI has access to. MCP servers flip this model. Era doesn't have a chatbot. Era has a data layer. Your AI of choice is the interface. This means you can ask about your finances wherever you already are — in the middle of a conversation about trip planning, while reviewing a contract, while building a budget spreadsheet. Your money shows up in context, not in a separate app. ## Security and trust Connecting your bank accounts to AI raises legitimate security questions. MCP addresses several by design: - **No credential sharing**: Your bank login credentials are handled by MX during authentication and are never stored by Era or seen by your AI. - **Scoped access**: Each MCP tool has defined inputs and outputs. Your AI can call "list transactions" but can't access raw database tables or internal systems. - **Revocable**: You can disconnect any AI client from your Era Context at any time, instantly cutting off access. Era adds additional protections on top of MCP: - AES-256 encryption at rest, TLS 1.3 in transit. - Your data is never sold, never used for advertising, and never shared without your explicit permission. - Every AI agent interaction requires explicit authorization. - A full activity log shows everything any AI agent has done on your behalf. ## Which AI clients support MCP MCP adoption is growing rapidly. Era works with any MCP-compatible client, including Claude, ChatGPT, Cursor, VS Code, GitHub Copilot, Gemini, Perplexity, OpenClaw, Manus, Cline, Hermes, and more. The list expands regularly as more AI tools adopt the standard. The important point is not any specific client — it's that you're not locked in. When a new AI client launches and supports MCP, it works with Era on day one. No integration needed, no feature request, no waiting. One protocol, every client. ## Why MCP matters more than any single AI model AI models improve rapidly. The model you use today will likely be surpassed within months. This creates a problem for any platform that couples its AI features to a specific model: every model upgrade becomes a migration, and users are stuck with whatever the platform ships until the next update. MCP decouples the data layer from the AI layer. Your financial data is accessible through a stable protocol regardless of which model is on the other end. When a new model launches — faster, cheaper, better at reasoning — you connect it to the same MCP server and get immediate benefits. No migration, no waiting for your finance app to integrate it, no re-training a chatbot. This is why Era built on MCP rather than embedding a specific model. The protocol outlasts any individual model generation. Your data stays structured and accessible, and the best AI available at any given moment can work with it. ## Getting started Era's Basic tier is free and includes two connected accounts with read-only MCP access. You can sign up at [era.app](https://era.app), connect a bank account, add one line of configuration to your AI client, and start asking questions about your money in natural language. The setup takes about five minutes. The shift in how you think about your finances takes a bit longer — but once your AI can actually see your money, you'll wonder why you ever managed it any other way. ### Introducing Era: AI-native personal finance via MCP Era is a personal finance platform that connects your bank accounts to any AI assistant via the Model Context Protocol (MCP). Instead of locking you into one chatbot, Era lets Claude, ChatGPT, Gemini, or any MCP-compatible agent read your financial data, generate insights, and trigger automations on your behalf. ## What is the Model Context Protocol? MCP is an open standard that lets AI agents securely access external data and tools through a uniform interface. Think of it as a USB-C port for AI: any compliant client plugs in, authenticates, and gets structured access to the host application's capabilities. Anthropic published the spec in late 2024, and adoption has grown rapidly across developer tooling, enterprise platforms, and now personal finance. ## How does Era use MCP? When you connect your bank accounts through Era, the platform aggregates balances, transactions, and account metadata from thousands of financial institutions. Era then exposes this data through a standards-compliant MCP server. Your AI agent of choice connects to that server, authenticates with your credentials, and gains read and write access to a curated set of financial tools. Available tools include querying transaction history, categorizing spending, creating automated money movement rules, and retrieving account balances. Every tool call is scoped by the permissions you grant — your agent cannot exceed the access level you define. ## Which AI agents work with Era? Any agent that implements the MCP client specification can connect. Today that includes Claude, ChatGPT, OpenClaw, Cursor, Manus, Gemini, and a growing list of compatible clients. Era also ships Agency, a first-party AI companion built directly into the app for users who prefer a turnkey experience. ## Is my financial data secure? Era uses bank-level AES-256 encryption for data at rest and TLS 1.3 for data in transit. All MCP connections require OAuth 2.1 authentication with scoped permissions. Your credentials are never stored on Era servers — your bank connections are handled through SOC 2 Type II-certified aggregation infrastructure. Every tool invocation is logged, auditable, and revocable from Era Context. ### How to add a manual account to Era You can add a manual account to Era in under two minutes — from Era Context directly, or by asking your AI to create it. Manual accounts sit alongside your connected bank accounts, show up in your spending views, and accept transactions individually or via CSV import. ## What is a manual account? A manual account is any financial account you track without a live bank connection. Cash, credit unions not yet in the bank sync network, loans you service offline, HSAs, or accounts you'd simply rather manage yourself — all of these belong here. Once added, a manual account works like any other account in Era Context: you can add transactions, import history, and include it in your financial overview. ## How do you add a manual account from Era Context? 1. Open Era Context and go to **Bank connections**. 2. Select **Add manual account**. 3. Enter a name (e.g., "Cash wallet"), select an account type (checking, savings, credit, loan, asset, or other), and confirm the currency. 4. Save. The account is immediately active and visible in your financial overview. That's it. You can start adding transactions right away. ## How do you add a manual account through your AI? If you have Era Context connected to an MCP-compatible client — Claude, ChatGPT, OpenClaw, or any other — you can create a manual account without opening the app. Ask your AI: "Add a manual account called 'Cash wallet' with a starting balance of $200." Your AI uses Era's tools to create it and confirms when it's live. You can log transactions the same way: "Add a $45 cash grocery purchase to my Cash wallet account from yesterday." This is useful when you're already in a conversation about your finances and want to record something without switching apps. ## How do you add transactions to a manual account? **One at a time from Era Context:** 1. Go to **Transactions** and select your manual account. 2. Select **Add manual transaction**. 3. Enter the date, description, and amount. 4. Save. **One at a time through your AI:** Ask your AI: "Add a $120 car repair charge to my Cash wallet account, dated today." **In bulk via CSV import:** 1. Export your transaction history from your bank or any other source as a CSV file. 2. Open the manual account in Era Context. 3. Select **Import transactions** and upload your CSV. 4. Confirm the import. Transactions appear immediately. You can also ask your AI to handle the import: "Import this CSV into my Cash wallet account." If any column mapping is ambiguous, your AI will ask you to clarify. ## How do you manage a manual account after setup? Once a manual account is live, you can: - **Rename or retype it** — change the account name or type from the account settings. - **Adjust the balance** — if the opening balance was wrong, update it directly. - **Toggle visibility** — hide a manual account from your overview without deleting it. - **Delete it** — remove the account and its transaction history if you no longer need it. Your AI can handle any of these: "Rename my Cash wallet account to 'Personal cash'" or "Hide my old loan account from my overview." ## What plans support manual accounts? Manual accounts are available on all plans, including the free Basic plan. Basic supports up to two accounts total (connected or manual). Organize ($9.99/month) and above support up to 15 accounts. See [pricing](/pricing) for a full breakdown. For a full picture of what Era Context can do once your accounts are set up, see [what to ask your AI about your money](/articles/what-to-ask-ai-about-your-money) and [how to automate your finances with AI](/articles/how-to-automate-finances-with-ai).