Anthropic launches Claude Personal Finance with read-only Plaid bank connections

Anthropic has launched Personal Finance, a beta feature that lets its Claude chatbot connect to users' bank and credit card accounts and answer questions about their money. The connection runs through Plaid, the widely used bank-linking platform, and access is read-only. The rollout began this week for Claude Pro, Max, and Ultra subscribers in the United States, according to The Verge and TechCrunch.

What Claude does with the data

Users can ask questions such as "where did my money go this month" and get spending totals sorted into categories like rent and utilities, food and drink, and shopping, The Verge reports. Follow-ups are also supported: whether a figure is higher than usual, which bills remain before payday, or whether a large purchase is affordable.

According to TechCrunch, the tool does more than summarize. It analyzes transactions across linked accounts to flag duplicate charges, unexpected fees, and steadily rising bills, and it can explain concepts like annual percentage rate while calculating how much interest a user paid over the past year.

Read-only, with a published scope

Anthropic says Claude cannot move, manage, or invest money, and that the feature "isn't designed to give investment advice."

What Claude sees is limited to each account's bank name, the account name and type, the last four digits of the account number, balances, credit limits, and up to two years of transaction history. Because Plaid brokers the connection, Claude never handles banking passwords, The Verge notes.

TechCrunch reports an additional privacy detail: conversations inside the Personal Finance hub are not used to train Anthropic's models. A dev.to analysis points out the boundary of that promise — it applies to chats within the hub, while anything shared with Claude elsewhere falls under normal privacy settings.

Anthropic is also positioning the business model as part of the pitch. In a blog post quoted by TechCrunch, the company says Personal Finance carries no ads or sponsored recommendations and does not push its own credit cards, loans, or savings accounts.

Late to a crowded field

Claude is following, not leading. OpenAI introduced comparable finance tools for ChatGPT in May, also built on Plaid, and TechCrunch notes Anthropic has been expanding Claude into verticals such as Claude for Legal and Claude for Financial Services this year.

A dev.to write-up adds context around the timing. It reports that TestingCatalog spotted a "Money" tab in Claude's iOS app in mid-September, that an October update showed the Plaid wiring, and that days before launch major US banks began blocking unauthorized AI agents found inside customer accounts. That sequence, the author argues, is the real lesson: banks shut down agents that logged in as customers, while permissioned, read-only integrations got a product page. The same post notes that Perplexity shipped a Plaid-based finance feature in April and that Plaid's network reaches more than 12,000 financial institutions, making Claude's arrival a distribution bet rather than a technical first.

Why it matters

The launch signals where AI assistants are heading: from generic financial advice to reading your actual transaction data. It also confirms the industry's emerging ground rules for AI touching money — read-only access, a published data scope, a permissioned aggregator in the middle, and a training carve-out for the sensitive conversations.

Anthropic itself frames this as a beginning. The Verge quotes the announcement describing Personal Finance as a first step, with bigger money decisions best served by Claude working alongside licensed professionals and regulated firms, and the company is inviting banks and wealth managers to help shape what comes next.

Two caveats are worth keeping in mind. The privacy carve-out stops at the hub's edge, so users should check their settings before pasting statements into ordinary chats. And a dev.to analysis quotes Anthropic's own warning that Claude can make mistakes, advising users to ask which transactions sit behind a number and to verify important figures with their bank.