Claude’s new Personal Finance tool aims to help analyze your spending – read-only, of course

Anthropic announced on Thursday that Claude can now connect to users' bank and credit card accounts and answer questions about their finances. The feature, called Personal Finance in Claude, is limited for now to Pro, Max and Ultra subscribers in the United States.
Connections are made through Plaid, the account-linking service widely used by budgeting and fintech apps, and access is read-only. Users can attach as many accounts as they like. For the accounts a user chooses to connect, Anthropic says Claude can see the bank's name, the account name, type and last four digits, balances and credit limits, and up to two years of transaction history.
Once accounts are linked, users can ask ordinary questions in plain language: where the money went this month, or whether a vacation in six months looks affordable. Claude can also build visual outputs. Anthropic told ZDNET that a user might request a weekly spending chart, a calendar of upcoming due dates, or a dashboard of chosen numbers, and Claude assembles it as an artifact that can be adjusted and revisited. Artifacts are Claude's persistent, editable output format, which makes a budget view something closer to a saved document than a one-off chat answer.
Privacy has been the central question around linking financial accounts to AI assistants, and Anthropic drew a boundary around the new mode. Conversations held inside the Personal Finance hub are not used to train its models, according to the company. That protection is tied to the hub itself: financial data or discussions that happen with Claude outside it fall under whatever the user's general Claude privacy settings allow.
The company also positioned the feature as a counterweight to its business model. In its announcement, Anthropic said Personal Finance carries no ads or sponsored recommendations and does not offer credit cards, loans or savings accounts of its own, arguing that money questions therefore get a straight answer. That contrasts with conventional personal finance apps and comparison sites, whose revenue often comes from product referrals.
On accuracy, Anthropic's claim is narrower than a guarantee of correctness. It says Claude works from actual transaction data rather than memory, so the feature should not fabricate figures, and it can list the exact transactions behind any number so a user can check the arithmetic. The company still acknowledges that, like any AI, Claude can make mistakes and recommends confirming anything important with the bank.
Anthropic framed the launch as a response to existing demand. Its April usage report found that financial advice was one of the main subjects users bring to Claude. The new mode packages that behavior into a dedicated surface rather than leaving it as improvised chat.
The feature mirrors OpenAI's ChatGPT Finances, which launched in May and likewise connects accounts. ZDNET reported that many Reddit users describe ChatGPT's version as accurate and reliable, while one poster compared its potential to established budgeting apps such as Mint or Monarch but found it weaker on historical syncing, high-quality dashboards and generating full statement PDFs. That comparison is a useful marker for what to watch: the hard part of a finance assistant is less the conversation than the plumbing behind it, including reliable history imports and exportable documents.
Anthropic is careful to frame the product as a starting point rather than an authority. It says bigger money decisions are best served by Claude working alongside licensed professionals and regulated firms, and that it is exploring how to make that possible inside Claude. It also invited banks, wealth managers and financial services companies to help shape the effort, which suggests a more integrated planning product may follow. A likely trade-off is that deeper integration with institutions would require heavier compliance, identity checks and data agreements than the current read-only Plaid link.
The natural-language interface is the main differentiator against traditional budgeting tools. Most finance apps present budgets through dashboards the user must interpret; a chat assistant answers a specific question directly and follows up. That is also the weakness, since a conversational answer is only as trustworthy as the data behind it, and users are being asked to extend trust to a model with a read-only view of their accounts.
Why it matters: Developers and product teams building anything in personal finance now have a second major AI assistant offering account-linked analysis, which raises the bar on conversational interfaces and on privacy messaging around sensitive data. The hub-specific no-training promise, and the fact that the same protection does not automatically extend outside the hub, is a design detail worth studying for anyone handling regulated user data in an assistant. Teams weighing similar features should also note Anthropic's own positioning. This suggests the near-term competitive ground is trustworthy read-only analysis and exportable artifacts, with deeper professional and institutional integration, in my inference, as the next contest.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
Anthropic has released what appears to be its answer to ChatGPT Finances.