MultiX0/last-archive

Implements a microservices architecture with a Go-based high-concurrency crawler, Python embedding service, and Node.js orchestration layer communicating over Docker's internal network, combined with Qdrant vector storage and SQLite metadata. Integrates Ollama for local LLM inference via an OpenAI-compatible Go bridge, enabling fully offline RAG with semantic search across archived content. Performance and accuracy are directly tied to crawl volume—the system requires substantial indexed data to generate meaningful responses.