Commit History

11 days ago

  1. Optimize retrieval latency: rerank top 20, deferred parent promotion, and real-time streaming Running

    bf1046c
  2. Fix Groq 413 TPM rate limit by budgeting prompt context and handling streaming errors

    30996e1

  1. Pin pyarrow==18.1.0 to resolve NumPy 2.0 conflict

    2929eee

  1. Update default Groq model to openai/gpt-oss-120b and fix Dockerfile reranker model

    923d3a8

  1. Bake AI models directly into Docker image to prevent timeouts

    eaac407

  1. Add 📈 emoji favicon to UI

    2bf7b60

  1. Updated README

    6b36f7a
  2. Fix requirements for FastAPI backend

    b840f9c
  3. Add Hugging Face configuration

    6096e37
  4. Track pkl and png with LFS

    d7b5c29
  5. Ready for production with LFS for PDFs

    66afdf1
  6. Initial commit

    454f118