We research, rank, and track Colosseum hackathon projects — then package the intelligence so Hermes and OpenClaw agents reach the signal before the market does.
The same database an agent reads. Search, filter, rank, and inspect every Colosseum project from one surface — these rows are /api/projects.
The methodology is visible to humans and packed into the skill layer for agents. Six axes — comparable across thousands of teams.
No mock widgets. The cards, ticker, ranking bars, filters, and skill export are driven by the live API — the chart below is the actual distribution of total scores across the database.
Projects are normalized from the Colosseum research source — names, teams, sites, socials, categories.
Product, team, branding, market and token signals become comparable metrics across thousands of entries.
Filters expose token-watch, strong-founder, best-branding, and under-radar tracks for analyst review.
The same database ships as /api/skill.md — an agent-readable layer for Hermes and OpenClaw.
The live database becomes a portable, agent-readable layer for Hermes and OpenClaw. The export is generated from the same API-backed rankings rendered on this page.