quivr
FreeDump all your files and chat with it using your generative AI second brain using LLMs & embeddings.
FreeFree tier
Inputs: file, textOutputs: text
About quivr
Quivr is an opinionated RAG (Retrieval-Augmented Generation) framework that helps developers integrate generative AI into their applications. It is designed to be fast, efficient, and customizable, supporting any LLM (OpenAI, Anthropic, Mistral, Gemma, etc.), any file type (PDF, TXT, Markdown, etc.), and allows adding internet search and other tools. The core can be installed via pip and used with just a few lines of code, making it easy to add a 'second brain' to any product. It also integrates with Megaparse for file ingestion. Quivr is fully open source and free to use.
Key Features
Opiniated RAG that is fast and efficient, allowing focus on product development
Works with any LLM including OpenAI, Anthropic, Mistral, Gemma, etc.
Works with any file type: PDF, TXT, Markdown, and supports custom parsers
Customizable RAG: add internet search, tools, and other integrations
Integration with Megaparse for advanced file ingestion
Simple installation via pip: pip install quivr-core
Easy integration into existing products with minimal code (5 lines of Python example)
Supports local models via Ollama and cloud APIs
Pros & Cons
Pros
- Free and open source with MIT license
- Works with a wide range of LLMs and local models
- Supports multiple file types and custom parsers
- Easy to integrate: can be added to a project with few lines of Python
- Customizable: allows adding internet search, tools, and workflows
- Active development on GitHub with community contributions
Cons
- Requires Python environment and some programming knowledge to set up
- Documentation is still evolving (some features marked as 'stay tuned')
- Relies on external API keys for cloud LLMs (e.g., OpenAI, Anthropic)
- No pre-built GUI or hosted service available by default (though Quivr.com exists separately)
Best For
Building a personal assistant or second brain to query personal filesAdding RAG-based Q&A capabilities to existing software productsIngesting and querying a variety of document formats (PDFs, text files, markdown)Rapid prototyping of custom knowledge bases with flexible LLM and vectorstore choicesEducation and research projects requiring a simple, open-source RAG framework
FAQ
What LLMs does Quivr support?
Quivr works with any LLM, including OpenAI, Anthropic, Mistral, Gemma, and local models via Ollama.
What file types can Quivr process?
Quivr works with any file type out of the box, including PDF, TXT, Markdown, and you can add your own parsers for custom formats.
Is Quivr free and open source?
Yes, Quivr is fully open source and free to use. The code is available on GitHub under an MIT license.
How do I install Quivr?
Install the Python package with pip: pip install quivr-core. Then use it with a few lines of code as shown in the README.
Can I integrate Quivr into my existing application?
Yes, Quivr is designed for easy integration. You can import the Brain class and add RAG capabilities to your product with minimal code.