The Problem
Solo founders waste weeks on market research before writing a single line of code. They spend hours on Google, reading blog posts, and guessing at market size — only to discover their idea has 15 competitors they didn't know about.
The Solution
GSR runs 5 parallel AI research agents that do the work in minutes:
- Researcher — validates pain signals across online communities
- Competitor Analyst — maps the competitive landscape and identifies gaps
- Market Sizer — estimates TAM/SAM/SOM with conservative methodology
- Value Skewer — analyzes value delivery using DeMarco's Value Array framework
- Judge — delivers an evidence-based BUILD/KILL verdict
How It Works
gsr idea # Interactive interview about your idea
gsr research # Run all 5 agents in parallel
gsr score # Get the final verdict
Each agent produces a structured markdown report. The Judge synthesizes everything and earns its BUILD verdict through argument — the default stance is KILL.
Key Design Decisions
- Default-kill stance — the Judge doesn't give participation trophies. If the evidence doesn't support building, it says so.
- Conservative market sizing — shows the math, doesn't inflate numbers to make founders feel good.
- Evidence over vibes — every claim needs data. Forum posts, competitor revenue, search volume — not "I think people would want this."