Haiku RAG

Example Use Cases

• 1

  Advanced document search with hybrid vector and full-text retrieval.

• 2

  Question answering with citations and multi-turn conversational memory.

• 3

  Complex research workflows and analytical tasks via code execution.

Description

Haiku RAG provides a powerful hybrid search engine combining vector and full-text search with advanced reranking and multi-agent research workflows. It supports conversational RAG with session memory, complex analytical tasks via sandboxed Python execution, and visual grounding of document chunks. The server is local-first with multiple embedding and QA providers, accessible via CLI, Python API, and MCP protocol for integration with AI assistants like Claude Desktop.