AlphaGo-inspired MCTS for document retrieval. No vectors, no embeddings, no chunking — just reasoning. Upload PDFs, ask questions, get answers with citations.

Code-aware search and navigation engine, powered by vectorless.

Vectorless, reasoning-based RAG that runs 100% on-device (Ollama + Qwen) — grounded answers with a 0–100 confidence score that abstains when unsure.

A vector-less RAG works fully in local NO internet needed, with webUI

Agentree: index PDFs into a hierarchical section tree for agent retrieval (Claude/Codex structured completions, not vector chunking).