A candidate-focused discovery and research platform for the Y Combinator ecosystem.
Status: Active development · Core features implemented · Research agent operational
- Explore the YC company directory (5,600+ companies) with rich filtering
- Search naturally ("W24 B2B infra startups, small team, remote")
- Research any company on-demand via an autonomous agent that crawls public sources and returns evidence-backed dossiers on product positioning, team/founder signals, hiring activity, funding, and market presence — with full source traceability and freshness timestamps
The agent runs as a background Trigger.dev task. It spins up an E2B Desktop sandbox with a persistent Chromium browser, then executes a multi-stage loop: search → crawl → analyze → extract → synthesize. All events stream to the frontend in real-time, including a live VNC view of the browser. Runs are budget-bounded (time, pages, domains) with diminishing-returns stopping rules and are cancellable at any time.
| Layer | Stack |
|---|---|
| Database | Neon Postgres + pgvector (hybrid semantic/lexical search) |
| Jobs | Trigger.dev v4 + E2B Desktop sandbox |
| AI | Anthropic Claude (primary), Gemini, OpenAI |
| Web Research | Serper, Playwright, Firecrawl |
| UI | Next.js, Linear-inspired design system |
Prerequisites: Bun 1.3.4+, Neon Postgres w/ pgvector, Trigger.dev account, API keys for E2B / Anthropic / Gemini / OpenAI / Firecrawl / Serper
bun install
cp env.example .env.local # fill in API keys + DB URL
bun run db:migrate
bun run dev
# In separate terminal:
bunx trigger.dev@latest devbun run dev / build / lint
bun run db:migrate / db:ingest-yc
bunx trigger.dev@latest dev / deployBuilt with ❤️ for the YC community