Everyone argues about whether AI kills SaaS.
We measure it.
SaaS Bounties is an empirical research project, not an opinion. Developers build AI-assisted solutions in real, well-defined SaaS categories and submit them for private review. We measure effort and quality, then publish only the anonymized metrics — never your source code. No hot takes. Just evidence.
What is this, exactly?
Every time a new AI coding tool drops, the same debate erupts: "SaaS is dead." "Founders are toast." "Why pay $50/month when Claude can just build it?" The counterargument follows immediately: "Distribution matters." "Switching costs are real." "Enterprise won't self-host." Both sides are arguing from intuition. Neither side has data.
SaaS Bounties is a structured experiment to generate that data. The question we're actually asking is narrow and measurable: how much time and effort does it take a developer using AI-assisted coding tools to build software that is functionally equivalent to a real SaaS product? We're not asking whether that software will ever compete in the market. We're just measuring replication effort.
Developers pick a bounty — a detailed feature specification derived from an existing SaaS product's public documentation — build the software using AI tools, and submit their work for private review. Submissions are never made public and are only reviewed to verify bounty eligibility — you retain full ownership of your code. We record the tools used, time spent, and whether the spec was met. Approved submissions earn bounty points and appear on the leaderboard. Only the aggregated, anonymized metrics are published — never your source code.
How it works
Pick a bounty
Browse the Bounty Board. Each bounty is a detailed feature spec for a real SaaS category — time tracking, expense management, and more.
Build with AI tools
Use Claude Code, Cursor, Copilot, or any AI-assisted coding tool to build a working implementation. The spec is downloadable as a markdown file to feed directly to your AI agent.
Submit your work
Upload your codebase and AI prompt history privately, note the tools you used and hours spent, and submit for review. Your code is never shared or made public — you keep full ownership.
Earn points
Approved submissions earn bounty points and appear on the leaderboard. Earn enough to unlock badges. Only anonymized metrics (time, tools, completion) are published — never your source code.
Scope matters. Here's ours.
This project has a tight scope. We're specific about it because it's the only way the data means anything.
What we measure
- Time and effort to replicate defined SaaS features
- Which AI tools were used and how
- Feasibility across different product categories
- Quality of AI-generated implementations
What we don't measure
- Whether replicas can compete in the market
- Distribution, trust, or switching costs
- Whether any SaaS company is at risk
- Whether AI agents replace the need for software
Ready to contribute?
Pick a bounty, build something real, and help settle the debate with data.