3BQRIIINO. INDEPENDENT TECH JOURNAL
← All stories

Anthropic / Claude / Personal Finance / AI News · November 6, 2026

Claude Can Now Read Your Bank Account — But It Can't Touch Your Money

Anthropic's Personal Finance lets Claude answer money questions from your real bank and card accounts, read-only through Plaid. It's a US-only beta for Pro, Max, and Ultra — and it arrives six months after ChatGPT did the same thing.

Claude Can Now Read Your Bank Account — But It Can't Touch Your Money

Anthropic announced Personal Finance in Claude on 5 November 2026, a beta that lets Claude look at your actual bank and card accounts and answer questions about your money in plain language — "Where did my money go this month?", "Can I afford a trip this spring?", "What's left before payday?" The connection runs through Plaid, the service that already links banks to finance apps, and it is read-only: Claude can read your balances and transactions, but it cannot move, manage, or invest a cent.

It is rolling out in beta this week to Pro, Max, and Ultra subscribers in the US only. Free, Team, and Enterprise plans do not get it, and there is no announced date for other countries — Anthropic says the Personal Finance page does not even appear outside the US, even on a VPN.

What Claude can see — and what it can't

The access list is short, and Anthropic publishes it exactly. For only the accounts you choose to connect, Claude sees:

You link up to 16 banks, including checking, credit card, loan, and investment accounts — though for investments, loans, and cards, Claude reads balances and limits only, not holdings, trades, interest rates, or loan terms, according to Anthropic's Help Center as reported. You sign in to your bank yourself through Plaid, so Claude never asks for or sees your bank password.

What it cannot do is the whole design decision. It cannot transfer money, pay a bill, cancel a subscription, or place a trade. It is not designed to give investment advice. And Anthropic leans hard on one promise: Personal Finance has no ads or sponsored recommendations, and the company offers no credit cards, loans, or savings accounts of its own to sell you. Conversations inside the Personal Finance hub are not used to train Anthropic's models; anything you share with Claude outside the hub falls under your normal privacy settings. You can unlink an account at any time, after which Claude can no longer see its data.

Anthropic's own example

The announcement walks through one worked example — the company's demonstration, not an independent test. Late in the month, with payday still 10 days away, a user asks "Where did my money go this month?" Claude totals $4,309.37 across the connected accounts, with more than half going to rent and utilities and $1,024.51 to food and drink. Follow-ups — "Is that more than usual?", "Anything else I'm missing?", "So what's left before payday?" — surface a membership renewal, four payments due out of the checking account before payday, and $1,393.22 left over.

Beyond the monthly totals, the feature is pitched as a way to spot what is off: ask whether anything looks strange and Claude reads across every connected account to point at duplicate charges, unexpected fees, or a bill that has crept up all year. Bigger questions work the same way — "Am I spending more than I make?" and "What's an APR, and what has mine cost me this year?" — with Claude explaining the term and then adding up the interest actually charged. You can also ask for a weekly spending chart, a calendar of what is due, or a dashboard of chosen numbers, which Claude builds and you keep adjusting.

The caveat Anthropic prints itself

Anthropic is unusually direct about the accuracy limits, and it prints the warning in the announcement itself: "Claude can make mistakes in its answers, so ask which transactions are behind a number, and check important figures with your bank."

The Help Center, as reported, names three specific ways a total can mislead: balances can be a day or two old; an account connected through two different logins can be counted twice if the bank presents it differently under each; and when Claude cannot read a linked bank, that bank's spending is left out of the figures — which can make the money left before payday look larger than it is. For a feature that answers questions with numbers about your money, that is the paragraph to remember.

Six months behind ChatGPT

Claude is not first here. Perplexity shipped its Plaid finance integration on 9 April 2026, and ChatGPT Finances launched on 15 May 2026 — also on Plaid, also read-only, at first for US Pro subscribers alone. Claude arrives roughly six months after ChatGPT, on the same connector, doing the same job: read your bank feed, answer questions about it.

The difference is positioning, not plumbing. ChatGPT's finance tools arrived as part of a broader push into shopping and money features; Anthropic frames Personal Finance as a deliberately narrow first step, with no ads, no house products, and an explicit invitation to banks, wealth managers, and financial services firms to help Claude work alongside licensed professionals on larger decisions. The launch also drew a public accusation from crypto investor Anthony Pompliano, who said Anthropic had sought inside information from his competing finance startup Silvia months earlier — an unproven claim from a rival founder, with no reported response from Anthropic.

Why it matters

A chatbot that can read your bank account is a line crossed, even read-only. Two years of transactions across up to 16 banks is a detailed map of a person's life — where they shop, what they pay for, when money gets tight. Anthropic's answer is restraint by design: a published list of exactly what Claude sees, no password ever touching Claude, no ads, no training on hub conversations, and an honest warning that the answers can be wrong.

That restraint is also the limit. The feature is US-only, paid-plans-only, and in beta — most readers cannot try it today, and nobody outside the US can. But the direction is clear: the major chatbots are all becoming money assistants now, and the competition is no longer about who can answer a question, but about whose answers about your own money you trust. Anthropic is betting that a short, published access list and a read-only promise are the way to earn that trust.

Sources and classification

Classification: Official Announcement (Anthropic, 5 November 2026). Launch facts, the access list, the worked example, and the privacy statements are from Anthropic's announcement, read first-hand; plan/region limits and Help Center details are corroborated by the reporting below. The ChatGPT and Perplexity comparison dates are from launch reporting. Checked 6 November 2026.

Prefer reading in Arabic? You can translate this article into Arabic with Google Translate.