Contributing to Agentic RAG

Thanks for your interest in contributing. This project is open source and we welcome patches, docs improvements, and feedback.

How to contribute

Bug reports and feature ideas: Open a GitHub Issue (replace with your repo URL). Describe what you did, what you expected, and what happened (and your env: OS, Python version, Docker if relevant).
Code and docs: Open a Pull Request. Keep changes focused; link any related issue. We’ll review and may ask for tweaks.

Clone and install

git clone https://github.com/your-username/local-ai-agent.git
cd local-ai-agent
uv sync   # or: pip install -e ".[dev]"

Environment

cp .env.example .env
# Set at least POSTGRES_PASSWORD. Use REQUIRE_AUTH=false for local dev if you prefer.

Run dependencies (e.g. Docker)

docker compose up -d postgres qdrant redis

Run the API

uv run uvicorn src.api.main:app --host 0.0.0.0 --port 8000

See docs/DEVELOPER_GUIDE.md for full setup, code layout, and how to extend the system.

Unit tests: uv run pytest tests/ -v --tb=short
Live validation (with stack running at http://localhost:8080):
uv run python -m scripts.validate_phase1_live … validate_phase6_live

Python: type hints where helpful; async for I/O-bound paths.
All data and search are tenant-scoped; never bypass tenant_id in DB or Qdrant.
New API routes: add tests or validation coverage where possible.

By contributing, you agree that your contributions will be licensed under the same license as the project (MIT). See LICENSE.