LLM-based agent with two goals.
1a) Surf the web effectively, starting with mostly text-based sites like reddit and hackernews. Produce digests. 1b) Explore and summarize complex code repos. 2) Efficiently utilize advances in fast and efficient prefix caching. 3) Produce embeddings and cached contexts that can be effectively reused at will for deeper dives.
pip install -r requirements.txtpython src/main.py