Anthropic has given its Claude chatbot a window into its users' wallets. The company announced Personal Finance in Claude on November 5, a new beta feature that lets paying subscribers in the United States link their bank and card accounts and ask Claude about their money in plain language — with one firm boundary: the chatbot can look, but it cannot touch.

The feature is rolling out gradually this week to subscribers on Claude's Pro, Max and Ultra plans. It works across Claude on the web, desktop and mobile apps, and connects to US banks through Plaid, the service that already links bank accounts to thousands of apps. Users sign in to their bank through Plaid, so Claude never asks for — or sees — a bank password.

What Claude can — and can't — see

For each account a user chooses to connect, Claude can see the bank's name, the account's name, type and last four digits, balances and credit limits, and up to two years of transaction history. Users can link up to 16 banks, and ask Claude where the month's money went, what bills are coming up, or how much is left before payday.

For investment, loan and credit-card accounts, the view is narrower: Claude reads balances and credit limits only, not holdings, trades, interest rates or loan terms. And the read-only line is absolute. Claude cannot move money, pay a bill, cancel a subscription or place a trade, and Anthropic says the feature is not designed to give investment advice.

Anthropic is also keeping money out of the feature's business model. Personal Finance carries no ads or sponsored recommendations, and the company says it will not offer its own credit cards, loans or savings accounts through it — in line with the ad-free stance it set out for Claude earlier this year.

The privacy guardrails

Anthropic has published exactly what Claude can see for each connected account, and the list is short — six items. Conversations inside the Personal Finance hub are not used to train Anthropic's models, and unlinking a bank stops Claude from reading that bank's accounts, though it does not erase information already used in earlier conversations.

There are limits users should know about. Anthropic warns that balances can be a day or two old, that an account connected through two different logins can be counted twice, and that spending at a bank Claude cannot read is simply left out of its figures — which can make the money remaining before payday look larger than it really is.

Why now

The launch lands as AI assistants race to become useful with the most personal data people hold. Reuters noted the rollout comes ahead of Anthropic's highly anticipated initial public offering, and the feature is the company's clearest move yet into everyday financial life — a space where trust, not capability, is the real test.

Anthropic's own caveat, printed in its announcement, sets the tone: "Claude can make mistakes in its answers, so ask which transactions are behind a number, and check important figures with your bank." For now, the feature is US-only, with no date given for other countries.