Metadata-Version: 2.5
Name: haiku.rag-slim
Version: 0.88.1
Summary: Local-first agentic RAG with citations - hybrid search, reranking and multimodal retrieval over your own documents, no database server required - Minimal dependencies
Project-URL: Homepage, https://ggozad.github.io/haiku.rag/
Project-URL: Documentation, https://ggozad.github.io/haiku.rag/
Project-URL: Repository, https://github.com/ggozad/haiku.rag
Project-URL: Issues, https://github.com/ggozad/haiku.rag/issues
Project-URL: Changelog, https://ggozad.github.io/haiku.rag/changelog/
Author-email: Yiorgis Gozadinos <ggozadinos@gmail.com>
License: MIT
License-File: LICENSE
Keywords: RAG,agentic-rag,citations,docling,document-ingestion,embeddings,hybrid-search,lancedb,mcp,mcp-server,multimodal-rag,pydantic-ai,reranking,vector-database
Classifier: Development Status :: 4 - Beta
Classifier: Environment :: Console
Classifier: Intended Audience :: Developers
Classifier: Operating System :: MacOS
Classifier: Operating System :: Microsoft :: Windows :: Windows 10
Classifier: Operating System :: Microsoft :: Windows :: Windows 11
Classifier: Operating System :: POSIX :: Linux
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Typing :: Typed
Requires-Python: >=3.12
Requires-Dist: docling-core==2.93.0
Requires-Dist: fastmcp<5.0.0,>=4.0.2
Requires-Dist: httpx>=0.28.1
Requires-Dist: jinja2>=3.1.0
Requires-Dist: lancedb==0.38.0
Requires-Dist: pathspec>=1.0.4
Requires-Dist: pyarrow<25,>=16
Requires-Dist: pydantic-ai-slim[ag-ui,logfire,openai]<3.0.0,>=2.40.0
Requires-Dist: pydantic-monty<0.0.24,>=0.0.23
Requires-Dist: pydantic>=2.12.5
Requires-Dist: pylance==11.0.0
Requires-Dist: pypdfium2>=5.0
Requires-Dist: python-dotenv>=1.2.2
Requires-Dist: pyyaml>=6.0.3
Requires-Dist: rich>=14.3.3
Requires-Dist: textual-image>=0.8.5
Requires-Dist: typer<0.27.0,>=0.21.0
Requires-Dist: watchfiles>=1.1.1
Requires-Dist: zstandard>=0.23.0; python_version < '3.14'
Provides-Extra: anthropic
Requires-Dist: pydantic-ai-slim[anthropic]; extra == 'anthropic'
Provides-Extra: bedrock
Requires-Dist: pydantic-ai-slim[bedrock]; extra == 'bedrock'
Provides-Extra: cohere
Requires-Dist: cohere>=5.21.1; extra == 'cohere'
Provides-Extra: cross-encoder
Requires-Dist: sentence-transformers<6.0.0,>=3.0.0; extra == 'cross-encoder'
Provides-Extra: docling
Requires-Dist: docling-ibm-models==4.0.1; extra == 'docling'
Requires-Dist: docling-parse==7.16.0; extra == 'docling'
Requires-Dist: docling==2.124.0; extra == 'docling'
Requires-Dist: opencv-python-headless<5.0.0.0,>=4.6.0.66; extra == 'docling'
Provides-Extra: google
Requires-Dist: pydantic-ai-slim[google]; extra == 'google'
Provides-Extra: groq
Requires-Dist: pydantic-ai-slim[groq]; extra == 'groq'
Provides-Extra: ingester
Requires-Dist: aiosqlite>=0.20; extra == 'ingester'
Requires-Dist: asyncpg>=0.29; extra == 'ingester'
Requires-Dist: fastapi>=0.125; extra == 'ingester'
Requires-Dist: obstore<0.10,>=0.9; extra == 'ingester'
Requires-Dist: sqlalchemy[asyncio]>=2.0; extra == 'ingester'
Requires-Dist: uvicorn[standard]>=0.32; extra == 'ingester'
Provides-Extra: jina
Requires-Dist: torch>=2.0.0; extra == 'jina'
Requires-Dist: transformers<6.0.0,>=4.40.0; extra == 'jina'
Provides-Extra: mistral
Requires-Dist: pydantic-ai-slim[mistral]; extra == 'mistral'
Provides-Extra: s3
Requires-Dist: obstore<0.10,>=0.9; extra == 's3'
Provides-Extra: tui
Requires-Dist: textual>=8.2.4; extra == 'tui'
Requires-Dist: tree-sitter-json>=0.24.8; extra == 'tui'
Requires-Dist: tree-sitter>=0.25.2; extra == 'tui'
Provides-Extra: vertexai
Requires-Dist: pydantic-ai-slim[google]; extra == 'vertexai'
Provides-Extra: voyageai
Requires-Dist: pydantic-ai-slim[voyageai]; extra == 'voyageai'
Provides-Extra: zeroentropy
Requires-Dist: zeroentropy>=0.1.0a11; extra == 'zeroentropy'
Description-Content-Type: text/markdown

# haiku.rag-slim

The core of [haiku.rag](https://github.com/ggozad/haiku.rag), agentic RAG on LanceDB, Pydantic AI and Docling, with every heavy dependency an optional extra.

Most users want [`haiku.rag`](https://pypi.org/project/haiku.rag/), which is this package with Docling, the VoyageAI and Cohere embedders, every reranker and the terminal UI.

## Installation

Python 3.12 or newer.

```bash
uv pip install haiku.rag-slim
uv pip install 'haiku.rag-slim[docling]'          # PDF, DOCX, PPTX, XLSX, HTML, LaTeX, email and image conversion
uv pip install 'haiku.rag-slim[docling,anthropic,cross-encoder]'
```

The core includes OpenAI and Ollama support, the MCP server and Logfire. Without `docling`, documents can be converted through docling-serve.

Extras: `docling`, `tui`, `voyageai`, `cohere`, `zeroentropy`, `cross-encoder`, `jina`, `s3`, `ingester`, and one per model provider: `anthropic`, `google`, `groq`, `mistral`, `bedrock`, `vertexai`. Ollama and any OpenAI-compatible endpoint need none. [Installation](https://ggozad.github.io/haiku.rag/installation/) lists what each provides.

## Documentation

Quick start, CLI, Python API and MCP setup: [ggozad.github.io/haiku.rag](https://ggozad.github.io/haiku.rag/).
