Bounty Fit

Inspiration

I enter Solana bounties and online hackathons from China on weeknights. Twice I spent an evening on an entry before noticing the prize only paid out to a rail I can't receive, or that AI-assisted work was banned. Bounty lists show the prize; nobody checks the rules against my situation.

What it does

Bounty Fit takes pasted rules text and my saved payout & eligibility profile (country, time zone, payout methods I can receive, whether I build with AI help) and returns a traffic-light verdict — GO, CHECK or NO-GO. Every finding quotes the exact sentence it came from: excluded countries or "only open to" lists, payout rails (USDC/crypto, PayPal, Stripe, bank wire, US-only ACH, check), AI policy, required deliverables, the top prize, and the deadline converted to my time zone with a countdown. Change the profile and the verdict recomputes live, so the same rules give different answers to different people.

Who it's for

Indie builders outside the US, especially in places where payout rails are the real constraint, who hunt bounties and hackathons and want to stop wasting evenings on prizes they can't receive.

How I built it

I started from an empty folder with the Devpost Learn Skill Pack (npx skills add challengepost/learn-ai-basics) and worked through 1-start -> 2-scope -> 3-prd -> 4-spec -> 5-build with my coding agent. The scope/PRD/spec are in devpost/. It's a single static page in plain HTML/CSS/JS with no dependencies and no API keys; the extractor runs the same in the browser and in Node's built-in test runner.

Challenges

Rules pages are written by lawyers. Testing against a real Devpost rules page exposed false positives: "the laws of the United States" and a W-9 tax sentence made the US look excluded, and "Stripe payout to a US bank account" looked receivable for anyone with a bank account. Each became a test case and a fix.

What I learned

Planning first paid off: the scope interview forced me to name the kernel (the personal profile), which is exactly what the first real-world bug threatened. Explainable rules with quoted evidence beat a black-box answer for something you have to trust.

What's next

URL fetch, batch-ranking a list of bounties, a local-LLM fallback for unusual phrasing, Chinese UI.

Built With

• python