Ray: The Open-Source AI Financial Advisor That Runs on Your Laptop
Most personal finance apps sell you a dashboard. Ray sells you a conversation — one that happens entirely on your own machine, against your own bank data, with no cloud in the middle.
Launched on Product Hunt on April 12, 2026, Ray is an MIT-licensed terminal app that connects to over 12,000 financial institutions via Plaid and answers questions about your money the way a good advisor would: grounded in your actual balances, transactions, and goals. It ranked #2 on Product Hunt's daily leaderboard — and for once, the top-ten hype is pointing at something worth installing.
What Ray Actually Does
Ray is not another expense tracker with a chatbot pasted on top. It is a conversational AI advisor that maintains a persistent profile of your financial life and evolves it over time. You don't open an app to see charts — you ask a question and get an answer that's aware of your last paycheck, your credit card balance, and the subscription you forgot you were paying for.
Under the hood, Ray ships with 30+ tools that query your real financial data. It looks things up, runs calculations, and projects your balance forward based on actual income and spending patterns rather than generic averages.
The pitch is simple: stop staring at pie charts and start asking direct questions about your money.
The Local-First Architecture
This is the part that matters. Ray runs entirely on your computer. No cloud account, no server-side storage, no analytics pipeline quietly fingerprinting your spending:
- Your financial history lives in an encrypted database on your hard drive
- Your name is scrubbed before any prompt reaches an LLM
- You bring your own API keys for the model of your choice
Compare that to the dominant personal-finance apps, which typically store your transactions on their servers, aggregate them for "insights," and rely on you trusting their privacy policy. Ray's architecture makes the privacy question moot: the data never leaves your disk.
Installation
Ray is a single npm install — it runs on macOS, Linux, and Windows:
npm install -g ray-finance
That gets you the free, self-hosted version. You'll need your own Plaid credentials and an API key for whichever model you want to drive the agent. If you'd rather skip the setup, the hosted Ray Pro tier is $10/month and bundles managed Plaid connectivity and built-in AI.
Behavior Score and Gamification
Ray includes a daily 0-100 behavior score with streaks and unlockable achievements. On paper that sounds gimmicky, but the framing is clever: the score rewards actions — paying down a balance, sticking to a category budget, automating a transfer — rather than outcomes you can't directly control, like portfolio returns. It turns financial discipline into a game with feedback loops short enough to matter.
| Feature | Ray (Free) | Ray Pro | Typical Competitor |
|---|---|---|---|
| License | MIT (open source) | MIT + managed service | Proprietary |
| Data location | Local disk, encrypted | Local + managed auth | Vendor cloud |
| Institutions | 12,000+ via Plaid | 12,000+ via Plaid | Varies |
| AI model | Bring your own key | Bundled | Fixed vendor |
| Price | $0 forever | $10/month | $0–$15/month |
Why This Matters Now
The personal-finance app category has been quietly consolidating around a small number of venture-backed dashboards, and the tradeoff is always the same: give up your transaction data, get pretty charts in return. Ray is betting that there's a real audience for the inverse trade — keep the data, get a tool that actually does something useful with it.
It also arrives at a moment when local-first is having a second wind. Between consumer-grade models that run on a MacBook and coding agents that operate inside your repo, the idea of sending every piece of sensitive context to a hosted service is starting to feel like a choice rather than an inevitability.
The Bottom Line
Ray is the first AI financial advisor that treats privacy as a design constraint rather than a marketing bullet point. It's open source, it runs locally, and it answers real questions about your money instead of rearranging your data into another grid of widgets. If you've ever hesitated to hand your entire transaction history to a startup in exchange for a chart you'll look at twice, Ray is the first thing worth trying.
Install it with one command, plug in your own keys, and ask it something specific. That's the whole pitch — and in a category that's spent a decade chasing the wrong UI, that's more than enough.
