AI-powered document search for PDF-heavy workflows. Built during the Mistral Worldwide Hackathon 2026.
Upload PDFs, ask questions in plain language, and get answers with page-level evidence — reviewable and traceable. Powered by PageIndex: Document Index for Vectorless, Reasoning-based RAG.
| Directory | Description |
|---|---|
srv/ |
Go backend API and indexing/retrieval pipeline |
web/ |
React app for upload, query, and evidence review |
marketing/ |
Astro marketing site |
email/ |
Hono + Cloudflare Worker for waitlist capture |
video/ |
Remotion project for video assets |
docs/ |
System-of-record documentation |
cp srv/.env.example srv/.env # then set MISTRAL_API_KEY
task setup # install dependencies
task dev # start API + web appThe web app runs at http://localhost:3000 and the API at http://localhost:2137.
Run task help for the full list of available commands (build, lint, test, codegen, etc.).
- PDF-only ingestion (no other file types yet).
- Single-user MVP — no authentication.
- Local filesystem storage, no encryption at rest.
- Evidence ranges are section/page-level, not text spans.
