Harden LLM prompt to stop small models parroting passages

Surface LLM provider/model/last-error on /health for diagnostics

Bump to 0.2.0 and expose version + llm_enabled on /health

Add LLM_PROVIDER preset (groq/gemini/openrouter) for the free LLM

Switch LLM answers to a free OpenAI-compatible API (Groq default)

Add optional Claude-generated answers with extractive fallback

Point live demo at the working Railway deployment

Fail HF sync loudly on manual dispatch when HF_TOKEN is missing

Match HF sync default to the created Space name (Docuask)

Add Railway deploy config for the API

Switch deploy to HF Static Space (frontend) + separate API

Default HF sync to Sri-28/docuask and fill in demo link

Add GitHub Action to auto-sync main to Hugging Face Space

Polish, tests & single-image deploy (Session 5)

Add data-collection & feedback layer (Session 4)

Add retrieval + chat loop (Session 3)

Add upload & parse flow (Session 2)

Add CI workflow: pytest backend + frontend build

Add DocuAsk skeleton (React + FastAPI full-stack scaffold)

Initial commit