Nearly 1 in 10 founders paid themselves nothing this year.
On November 5, Anthropic launched Personal Finance in Claude: a read-only tool that connects to your bank and card accounts and answers questions about your money in plain language.
For a founder who pays the team before paying themselves, that matters more than for almost anyone, because your personal burn rate decides how long you can keep building.
This guide covers what the tool does, seven ways to use it to live lean without guessing, and the exact prompts to copy.
Below is the full breakdown, including how to run the same playbook today if you're outside the US.
Inside you'll find:
What Claude Personal Finance Actually Does (and What It Won't)
Your Startup Has a Runway. So Do You.
The 7-Play Playbook, With Prompts
Outside the US? Run the Same Playbook Today
The Fine Print, and Your Next Step
1. What Claude Personal Finance Actually Does (and What It Won't)
It doesn't manage your money. It reads it.
Personal Finance in Claude connects your bank and credit card accounts through Plaid, the same infrastructure behind many finance apps. You log in to your bank through Plaid, so Claude never asks for or sees your bank password. Access is read-only: Claude cannot move, manage or invest your money.
Once you connect your accounts (up to 16), Claude sees the bank and account names, account types, the last four digits, balances, credit limits and up to two years of transactions. Then you ask questions the way you'd ask a friend who happens to be good with spreadsheets.
According to Anthropic's launch details, it can:
Total your spending by category and compare it with earlier months ("Where did my money go this month?" followed by "Is that more than usual?")
List the payments due before your next payday
Flag duplicate charges, unexpected fees, subscriptions and bills that keep creeping up
Explain terms like APR and calculate how much interest you paid over the past year
Build charts, calendars and dashboards from your own data
→ The detail founders should notice is the two years of history. That's enough to see your real cost of living, not the version you remember.
What it won't do: give investment advice, sell you a credit card or a loan, or show you ads. Anthropic says the hub has no sponsored recommendations, and conversations inside it aren't used to train its models or added to Claude's shared memory.
Availability is the catch. It's a beta, rolling out this week to Pro, Max and Ultra subscribers in the United States only. It also arrives about six months after OpenAI launched a similar feature in ChatGPT in May, so this is now a category, not an experiment.
2. Your Startup Has a Runway. So Do You.
Most founders know the company's runway to the week. Very few know their own.
That's a problem, because in the early years they're the same number. When you pay your engineers a market salary and yourself a fraction of one, the company's survival depends on how long you can live on less.
The data shows how common that trade is. In Pilot's 2026 Founder Salary Report, based on 1,623 founders, those running teams of fewer than five employees paid themselves $75K on average. Founder pay only climbs as the team grows: $112K at six to ten employees, $156K past eleven.
The most telling detail is how founders cover the gap. 77% rely on at least one other source of income: savings, investments, a partner's earnings or consulting. For many founders, the real "salary" is a slow draw-down of personal savings.
→ Living lean isn't a sacrifice you make for the company. It's a strategy that buys the company time.
Paul Graham made this case back in 2009 in his essay "Ramen Profitable": once a startup makes just enough to cover the founders' living expenses, it's no longer at the mercy of investors. His example was two young founders who could break even on about $3,000 a month. The lower your personal number, the sooner you reach that point, and the better the terms you can negotiate when you do raise.
Airbnb is the famous version of what happens when founders fund the gap with debt. In 2008, Brian Chesky was about $25,000 into credit card debt keeping himself and the company alive, and Joe Gebbia owed tens of thousands more. They kept their credit cards in binders made for baseball cards. They dug out partly by selling Obama O's and Cap'n McCain's cereal at $40 a box, which raised about $30,000.
The lesson isn't "sell cereal." It's that a founder's personal finances are part of the company's risk. Until this week, understanding yours took a spreadsheet and a free Sunday. Now it takes a question.
3. The 7-Play Playbook, With Prompts
Each play follows the same structure: the situation, what Claude Personal Finance does, the prompt to copy, and the decision it should lead to. Run them in order the first time. After that, plays 3 and 6 become monthly habits.
The prompts are written for the Personal Finance hub, where Claude already sees your connected accounts. Replace anything in [brackets] with your own numbers. If you're working from a bank export instead, section 4 gives you the one line to add.
Play 1: Can I Quit My Job?
For: the person with the idea who hasn't jumped yet.
The situation. You have savings and an idea. What you don't have is an honest number for how long the savings last once the paycheck stops. Most people estimate their spending from memory, and memory is generous.
What Claude does. It reads your real transactions, separates what you'll keep paying from what disappears when you leave a job (commuting, work lunches), and flags what appears (health insurance, the tools your employer used to pay for).
Copy this prompt:
I'm considering leaving my job to start a company. Using my last 12 months of transactions, calculate my average monthly spending, split into fixed essentials (housing, utilities, insurance, debt payments, groceries, transport) and everything else. Then estimate how many months my current checking and savings balances would last in two scenarios: (1) my current lifestyle, (2) essentials only. Flag expenses that would likely disappear or appear once I stop working, such as commuting or employer health insurance. Show the result as a simple table.
The decision. The essentials-only number is your real ceiling. If it's shorter than the time you think it takes to reach revenue or a first round, you have three levers before quitting: save more, cut further, or line up part-time income. There's no prize for doing it without a safety net: 77% of founders already rely on some other source of income.
Play 2: Find Your Minimum Number
For: founders at pre-seed or seed deciding what to pay themselves.
The situation. After a raise, the salary question arrives fast, and it usually gets answered with a guess. Pilot found that 1 in 3 founders set their salary based on what the startup can afford, not on what they need or what the market pays.
What Claude does. It calculates your floor: the monthly amount below which your life stops working. That number, plus taxes, is the lowest salary that doesn't quietly drain your savings.
Copy this prompt:
Based on my last 12 months of spending, calculate my minimum sustainable monthly budget using real averages for rent or mortgage, utilities, insurance, minimum debt payments, groceries, transport and healthcare. Exclude discretionary spending. Then rank my discretionary categories by how much I spent on each, and show my monthly total if I cut the top three by 50%. Give me three numbers: my floor, my lean budget and my current budget.
The decision. Your floor is the least you can pay yourself without burning savings. Your current budget is the ceiling. Then compare both with the market, because the data says the method matters: founders who benchmark against market rates paid themselves $102K on average, against $79K for those who paid themselves whatever the startup could afford.
For reference, Kruze Consulting's 2026 payroll data puts the seed-stage CEO salary benchmark at $153K. You don't need to pay yourself that. But knowing your floor and the market lets you defend a number to your board instead of apologizing for one.
→ Pay yourself too little and you're quietly funding the company with your savings. Pay yourself too much and you shorten the runway. Your floor tells you where "too little" starts.
Play 3: The Leak Audit
For: everyone.
The situation. Cutting costs rarely starts with big decisions. It starts with the small charges nobody reviews: the free trial that converted, the bill that rose twice this year, the bank fee nobody noticed. Founders collect these faster than most, because trying new tools is part of the job.
What Claude does. This is the use case Anthropic highlighted most: flagging duplicate charges, unexpected fees, subscriptions and bills that keep rising.
Copy this prompt:
Review my last 6 months of transactions and list: (1) every recurring charge or subscription, with its monthly cost and when it started; (2) any bill that increased during the period, with the old and new amount; (3) any fee, interest charge or possible duplicate charge; (4) software subscriptions that look work-related. Sort everything by annual cost, highest first. Then ask me which ones I want to keep, and calculate the annual savings if I cancel the rest.
The decision. For each item, one of three answers: cancel, downgrade, or move it to the company. The third answer leads straight to play 4.
Play 4: Untangle Personal and Company Money
For: everyone, especially before a fundraise.
The situation. In the early months, the company card arrives late and the founder's card fills the gap: domains, cloud bills, a flight to see a customer, the first design tool. Each charge is small. Together they're an untracked loan from you to the company, and a mess for whoever does the books. Airbnb's founders show how far that loan can grow.
What Claude does. It finds the transactions in your personal accounts that look like business expenses and turns them into a list you can reimburse or hand to your accountant.
Copy this prompt:
Go through my transactions from [start date] to today and identify charges that look like business expenses for my startup: software and SaaS, cloud hosting, domains, business travel, coworking, contractors, advertising and professional services. Put them in a table with date, merchant, amount and a suggested category. Add a subtotal by category and a grand total. Mark anything you're unsure about so I can confirm it.
The decision. Either the company reimburses you, with documentation, or you record the total as money you've lent the company. Both are better than numbers that only live in your head. Ask your accountant which applies where you are, because the rules differ by country.
Play 5: Stress-Test Your Raise
For: founders raising now or within the next year.
The situation. You planned your personal finances around closing the round in three months. Rounds rarely respect plans. Paul Graham described raising as terribly distracting and noted that it can last for months. And the money is tighter: in Pilot's 2026 survey, the average amount raised nearly halved, from $7M to $3.9M, while the median stayed flat at $500K.
What Claude does. It combines your real spending, your balances and your upcoming bills to show how your personal runway holds up if the round slips.
Copy this prompt:
My startup currently pays me [amount] a month. If my fundraise is delayed and I have to cut my salary to [lower amount, or $0] starting next month, show how many months my savings last at my current spending and at essentials-only spending. Then tell me how many months of delay I can absorb in each case before my savings fall below [safety buffer, e.g. 3 months of essentials]. Present it as a table and a simple chart.
The decision. If a six-month delay breaks your household, you'll know before any investor does. You can cut costs now, on your terms, instead of accepting worse terms later because you need the money. That's Graham's argument about ramen profitability, applied to your own kitchen table.
Play 6: The 20-Minute Monthly Ritual
For: everyone, once the first five plays are done.
The situation. Plays 1 to 5 are a one-time cleanup. Without a habit, costs drift back within a quarter.
What Claude does. It runs the same review every month from a single prompt, comparing against previous months and the lean budget you set in play 2.
Copy this prompt:
Monthly review. Compare this month's spending with my average for the previous 3 months, by category. My lean budget is [amount from play 2]. Tell me: (1) whether I stayed under it and by how much; (2) the three categories that moved most, up or down; (3) any new subscriptions, fees or unusual charges; (4) payments due before my next income date; (5) how many months of essentials-only runway my current balances cover. Keep it to one screen.
The decision. One adjustment per month, no more. The goal is a routine you'll actually keep, not a monthly crisis. Block 20 minutes in your calendar on the first working day of every month and run this prompt.
Play 7: Know When to Call a Human
For: everyone.
The situation. Once your numbers are clear, bigger questions show up: whether to use savings or debt to extend your runway, how to structure founder pay for taxes, what to do with equity if the company does well.
What Claude does, and doesn't. Personal Finance isn't designed to give investment advice, and Anthropic itself warns that Claude can make mistakes and recommends checking important figures with your bank. The company has said it wants to explore working with licensed professionals for bigger money decisions. Until then, use Claude to prepare for that conversation, not to replace it.
Copy this prompt:
I'm meeting an accountant or financial advisor. Based on my accounts and our earlier analysis, write a one-page summary of my situation: income sources, monthly spending, debts with their interest rates, savings and personal runway. Then list the 8 most important questions I should ask them as a founder taking a reduced salary. Don't give me recommendations; help me ask better questions.
The decision. An hour with a professional who receives a clear one-page brief goes much further than an hour spent reconstructing your situation from memory.
4. Outside the US? Run the Same Playbook Today
Personal Finance is a US-only beta for now, and a large share of founders reading this are in Europe or Latin America. The good news: none of the seven plays depends on the live bank connection. They depend on your transaction history, and you already have that.
Option 1: export a CSV. Almost every online bank lets you download your transactions as a CSV or Excel file. Download 12 months from each account, remove account numbers if you prefer, and upload the files to a regular Claude chat. Then start any prompt above with this line:
Copy this prompt:
I've attached my bank and card transactions as files covering [months]. My current balances are: [account: amount]. Use only these files for the analysis, and tell me if any month looks incomplete or duplicated.
You lose the automatic updates and the dedicated hub. You keep the analysis. For the monthly ritual in play 6, export just the last month each time. One honest caveat: the hub's specific commitments (no training, nothing added to memory) apply to Personal Finance. A regular chat follows your account's privacy settings, so check them before uploading financial data.
Option 2: use a connector. If you'd rather have a live feed, Claude's connector directory includes PocketSmith, a personal finance app that says it's used in 190 countries. It offers a read-only version, so Claude can read your data without changing it. Bank-feed coverage depends on your country and PocketSmith plan, so confirm your bank is supported before you commit.
→ Outside the US, the feature you're waiting for is a convenience. The playbook works today.
5. The Fine Print, and Your Next Step
Three things to know before you trust the numbers:
The data can lag. Balances can be up to two days old, and accounts connected through more than one login can be counted twice. Check important figures with your bank, as Anthropic itself recommends.
It's a beta. Expect rough edges, and treat every total as a draft until you've verified the big ones.
The launch has critics. Investor Anthony Pompliano publicly attacked it, accusing Anthropic of seeking proprietary information from his personal finance startup, Silvia, months before launch. It doesn't change how the tool works for you, but it's part of the story.
None of this makes the decisions for you. It makes sure you take them with real numbers instead of memory and hope.
Your next step: download the Founder Personal Runway Kit, the free Excel template that goes with this guide. It includes a personal runway calculator, a minimum-number worksheet, a fundraise stress test, a reimbursement log for play 4, the monthly ritual checklist, and all seven prompts in both versions (Personal Finance hub and CSV). Fill it in once with the numbers from plays 1 to 5, and update it every month.
Your company's runway gets reviewed at every board meeting. Give yours 20 minutes a month.
