# HoundScraper documentation

Local-first content extraction and durable crawling for AI agents. Independently implemented under Apache 2.0.

Official source: https://ops.warhoundai.com/docs/houndscraper

## Start with the core

Run the API, PostgreSQL and React dashboard using Docker Compose. Core HTTP scraping, deterministic formats and persistent jobs require no model-provider account.

```
git clone https://github.com/warhoundai/HoundScraper.git
cd HoundScraper
python scripts/configure.py
docker compose up --build -d
```

## Know the release boundary

Version 0.1.0 is the stable core release. The 0.2 preview adds bounded agents, native Gemini/Vertex adapters, S3 storage, shared origin pacing and document/media processing. PostgreSQL, Chromium, scoped AWS S3, PDF/OCR, Office and human-speech fixtures are qualified separately. Live Vertex and search-provider acceptance remain outstanding. This is not full Firecrawl parity.