Your chatbot can now read your bank statements. On Thursday, November 5, 2026, Anthropic announced Personal Finance in Claude, a new feature that lets subscribers connect their bank and card accounts and ask Claude questions about their money in plain language — where it went, what bills are coming, and whether they can afford something. The access is read-only, which is the whole point of the design: Claude can see your money, but it cannot move a dollar of it. The beta is rolling out this week to paying subscribers in the United States, and it arrives just as Anthropic’s IPO is dominating the AI news cycle.
What Anthropic announced on November 5
Personal Finance is a new hub inside Claude. Link the accounts you choose, and Claude answers money questions using your own balances and transactions, in everyday language. TechCrunch reports that users connect bank and credit card accounts through Plaid, the service that links apps to banks, and that Anthropic describes the access as read-only: “Personal Finance can’t move, manage, or invest your money.”
The launch is deliberately narrow. It is a beta for individual subscribers on Claude’s Pro, Max and Ultra plans in the United States — Pro at $20 a month ($17 with annual billing), Max at $100, and Ultra at $200 — rolling out gradually, so some eligible subscribers will not see it yet. Free, Team and Enterprise plans are excluded, and only US banks can be linked. The feature works on Claude’s web app, desktop app, iOS and Android. Outside the United States, the Personal Finance section does not appear at all, even when connecting through a VPN, and Anthropic has given no date for other countries.
There is no extra charge for the feature itself, and Anthropic says it carries no ads or sponsored recommendations, and that it does not offer credit cards, loans or savings accounts of its own. That matters, because a finance tool that also sells financial products would have an obvious conflict of interest. The launch also lands in a telling week for Anthropic: Reuters noted the rollout comes ahead of the company’s highly anticipated IPO, the same offering covered in our report on Anthropic’s IPO.
How linking your bank actually works
The connection runs through Plaid, and the design keeps your bank password away from Claude entirely. You sign in to your bank inside Plaid’s own window, pick the accounts to share, and approve read-only access. Claude never asks for or sees your password. You can link up to 16 banks, and you can unlink a bank at any time to stop future reads.
That plumbing is not new — it is the same approach other AI assistants already use. But it is worth stating plainly, because the scariest version of this feature is one where you hand a chatbot your banking credentials. That is not what happens here. The bank login stays between you, your bank, and Plaid; Claude only receives the data you approved. If you want a practical guide to AI agents and how they handle permissions and tools, the same principle applies across the agent world: what an agent can read and what it can do are two different lists, and the second one should be much shorter than the first.
What Claude can see — and what it can’t touch
Anthropic publishes exactly what Claude receives for each connected account, and the list is short enough to quote in full: the bank’s name; the account name, type and last four digits; balances and credit limits; and up to two years of transaction history, including amounts, dates, merchants and categories.
What it cannot do is just as explicit. Claude cannot transfer money, pay a bill, cancel a subscription, or place an investment trade. It cannot manage your accounts or invest on your behalf. And Anthropic says the feature “isn’t designed to give investment advice” — it will explain your spending, not tell you where to put your money. The company also prints its own caveat in the announcement: “Claude can make mistakes in its answers, so ask which transactions are behind a number, and check important figures with your bank.” That is honest, and it is also the right way to use any AI that summarizes your finances: treat its answer as a starting point, and verify the number that matters before acting on it.
What you can actually ask it
The feature is built for everyday money questions rather than financial advice. Anthropic gives these examples of what it can do:
- Summarize spending by category — rent and utilities, shopping, dining — and compare one month with another
- Track upcoming bills and build a bill calendar so due dates stop sneaking up on you
- Flag duplicate charges, unexpected fees and out-of-the-ordinary purchases
- Explain interest payments and break down where a balance is going
- Build charts of your cash flow from the accounts you connected
It is the same shift happening across consumer AI this month — Amazon’s text-to-shop agent launched the same week, letting people buy through a chat message instead of a checkout page. The difference is the stakes: a shopping agent moves money by design, while a finance assistant is useful precisely because it does not.
The privacy promises, and the fine print
The feature’s trust case rests on three promises from Anthropic. First, conversations inside the Personal Finance hub are not used to train Anthropic’s models, and they are not added to Claude’s shared memory — so your spending does not become training data or leak into other chats. Second, there are no ads and no sponsored recommendations in the experience. Third, the read-only boundary is enforced at the connection level, not just promised in a blog post: Plaid describes the connection as read-only too.
The fine print is still worth reading. Your financial data and the answers Claude generates from it still pass through Claude, and they can remain in chats, artifacts or shared links you create. Unlinking a bank stops future reads, but it does not erase what was already discussed. And the two-year transaction window is a lot of history: a full picture of your spending, your merchants and your habits, sitting inside a chat product. That is a reasonable trade if you want an assistant that actually knows your money — but it is a trade, and it should be a conscious one. Link only the accounts you want analyzed, check what Claude says against your bank before acting on a figure, and remember that a beta is, by definition, still being tested.
Who gets access, and how to check
For now, access is limited to individual subscribers in the United States on Claude’s Pro, Max or Ultra plans. If you pay for one of those and live in the US, the Personal Finance section should appear in Claude’s menu once the rollout reaches your account — there is nothing to install, and no waitlist mentioned in the launch.
MacRumors notes that OpenAI added a similar finance feature to ChatGPT back in May 2026, and Perplexity shipped its own Plaid finance integration in April — Claude arrives roughly six months after ChatGPT, on the same connector, doing the same job.
There is also a controversy attached to the launch. Decrypt reports that crypto investor Anthony Pompliano accused Anthropic of trying to get inside information from his competing startup, Silvia, a personal AI chief financial officer, months before the rollout. Anthropic has not publicly addressed the accusation in the coverage so far, and the claim itself is unverified — but it underlines how crowded this corner of AI has become. Everyone can see the same prize: the assistant that knows your money becomes the assistant you ask about everything else.
The chatbot that reads your bank statements is here. Whether it becomes the way you manage your money depends on what it does with that access in the next few months — and on whether you trust the read-only line to hold.
