PY

haiku.rag

by ggozad/haiku.rag

0 views

Retrieval Augmented Generation based on SQLite

pythonsearchdatabaseDocument Processing

Haiku SQLite RAG

Retrieval-Augmented Generation (RAG) library on SQLite.

haiku.rag is a Retrieval-Augmented Generation (RAG) library built to work on SQLite alone without the need for external vector databases. It uses sqlite-vec for storing the embeddings and performs semantic (vector) search as well as full-text search combined through Reciprocal Rank Fusion. Both open-source (Ollama) as well as commercial (OpenAI, VoyageAI) embedding providers are supported.

Features

  • Local SQLite: No external servers required
  • Multiple embedding providers: Ollama, VoyageAI, OpenAI
  • Hybrid search: Vector + full-text search with Reciprocal Rank Fusion
  • File monitoring: Auto-index files when run as server
  • 40+ file formats: PDF, DOCX, HTML, Markdown, audio, URLs
  • MCP server: Expose as tools for AI assistants
  • CLI & Python API: Use from command line or Python

Quick Start

# Install
uv pip install haiku.rag

# Add documents
haiku-rag add "Your content here"
haiku-rag add-src document.pdf

# Search
haiku-rag search "query"

# Start server with file monitoring
export MONITOR_DIRECTORIES="/path/to/docs"
haiku-rag serve

Python Usage

from haiku.rag.client import HaikuRAG

async with HaikuRAG("database.db") as client:
    # Add document
    doc = await client.create_document("Your content")

    # Search
    results = await client.search("query")
    for chunk, score in results:
        print(f"{score:.3f}: {chunk.content}")

MCP Server

Use with AI assistants like Claude Desktop:

haiku-rag serve --stdio

Provides tools for document management and search directly in your AI assistant.

Documentation

Full documentation at: https://ggozad.github.io/haiku.rag/

Install

{
  "mcpServers": {
    "haikurag": {
      "command": "haiku-rag",
      "args": [
        "serve",
        "--stdio"
      ]
    }
  }
}
For more configuration details, refer to the content on the left

Related

Related projects feature coming soon

Will recommend related projects based on sub-categories