Clone GitHub repositories, embed it into PostgreSQL (pgvector), and ask an AI about it.
Built on LangChain + OpenAI, with future browser support to be able interactively ask questions on the web.
-
Install dependencies
python -m venv .venv .venv\Scrip 56DE ts\activate pip install -e .
This installs the dependencies and puts a
loreravencommand on your PATH. -
Configure
copy .env.example .env
Then edit
.envand setOPENAI_API_KEYandLORERAVEN_DATABASE_URL. -
PostgreSQL with pgvector
You need a Postgres database with the
pgvectorextension available. The quickest way is Docker:docker run -d --name loreraven-pg -p 5432:5432 \ -e POSTGRES_PASSWORD=postgres -e POSTGRES_DB=loreraven \ pgvector/pgvector:pg16
LoreRaven creates the extension and its table automatically on first ingest.
loreraven ingest https://github.com/pallets/flask
loreraven chat