Grounded Document Agent: cited PDF Q&A with LlamaParse, LlamaIndex and local Ollama
Roger_M_Taylor · x · 2026-09-22
An open-source Grounded Document Agent (3.7k GitHub stars) answers questions on long PDFs with numbered citations traceable to exact pages. The author argues many failures happen at parsing — flattened tables, lost chart structure, broken reading order — before retrieval even matters. Stack: LlamaParse for layout-aware Markdown with page metadata, LlamaIndex for indexing/retrieval, Ollama running local qwen3:4b-instruct.
More from coding & agent
- Agent memory design question: what deserves a place in permanent memory? — Miserable-Shock3552 · 2026-09-22
- Dev demos near real-time editing with Gemini Flash Lite, plans open-source Chrome extension — Saboo_Shubham_ · 2026-09-22
- Local MLX vs hosted API: two AI agents battle in Chrome's T-Rex game — usamawahabkhan · 2026-09-22
- A silent watcher race dropped agent logs for a month: full post-mortem of a tailing bug — ClaudeCdGuy · 2026-09-22
- D-RAC cuts enterprise document chunking tokens by 95.7%, cost by up to 85.6% — Yellow-AI-NLP · 2026-09-22
- Did the Action Happen? Designing Confirmation Logic for Agent Tool Calls That Time Out — felix_baron · 2026-09-22