multivector-search

🐊 Snappy's unique approach unifies vision-language late interaction with structured OCR for region-level knowledge retrieval. Like the project? Drop a star! ⭐

High-performance late-interaction retrieval engine for on-prem AI. ColBERT/ColPali multi-vector search with Rust fused MaxSim, Triton GPU kernels, ROQ quantization, LEMUR routing, WAL-backed CRUD, and a FastAPI server — single machine, CPU or GPU.

Vision-first document RAG for PDF, DOCX, and image QA/extraction with ColQwen2, Qwen2.5-VL, Qdrant, and FastAPI.

Multivector Windows matrix engineered to generate concentrated force fields for TikTok video visibility in 2026. Force distribution, matrix stability, no authentication required.

Conversational movie recommendation system powered by RAG architecture with semantic search and multi-vector embeddings.