Neura Market
Quiver logo

Quiver

Free

Dump your brain into your GenerativeAI Vault ![GitHub Repo stars](https://img.shields.io/github/stars/StanGirard/quiver?style=social)

FreeFree tier
Outputs: text
Type
Open Source
Company
QuivrHQ
Visit Website GitHub (40 stars)

About Quiver

Quivr is an opinionated, open-source RAG (Retrieval-Augmented Generation) framework designed to help developers integrate generative AI into their applications quickly and efficiently. It supports any LLM (OpenAI, Anthropic, Mistral, Groq, Llama, and local models via Ollama), any vector store (PGVector, Faiss), and any file type (PDF, TXT, Markdown, etc.). Quivr allows customization of the RAG pipeline, including adding internet search and tools. It integrates with Megaparse for enhanced file ingestion. Installation is simple via pip ('pip install quivr-core'), and a basic RAG can be created with just five lines of code. Quivr is the core brain behind Quivr.com and is actively developed with ongoing improvements.

Key Features

Opinionated and efficient RAG pipeline
Works with any LLM (GPT4, Groq, Llama, etc.)
Supports any vector store (PGVector, Faiss)
Accepts any file type (PDF, TXT, Markdown, etc.)
Customizable RAG with internet search and tool integration
Seamless integration with Megaparse for file ingestion
Simple installation via pip and 5-line setup
Supports local models via Ollama

Pros & Cons

Pros
  • Free and open-source
  • Opinionated design reduces RAG complexity
  • Supports a wide variety of LLMs and vector stores
  • Easily customizable with additional tools and parsers
  • Quick setup with minimal code required
Cons
  • Requires Python and API key configuration for remote LLMs
  • Opinionated approach may not suit highly specialized RAG needs
  • Still in active development; some features are planned but not yet available
  • Documentation and community resources may still be growing

Best For

Building a personal AI assistant or second brainIntegrating GenAI into existing products with minimal RAG overheadCreating file-based question-answering systems (PDFs, text files, etc.)Developing customizable RAG workflows for enterprise applications

FAQ

What is Quivr?
Quivr is an open-source, opinionated RAG (Retrieval-Augmented Generation) framework that helps developers build generative AI applications. It acts as a second brain by ingesting files and answering questions based on their content.
How do I install Quivr?
Install Quivr using pip: 'pip install quivr-core'. Requires Python 3.10 or newer.
Which LLMs does Quivr support?
Quivr works with any LLM, including OpenAI, Anthropic, Mistral, Google Gemma, Groq, Llama, and local models through Ollama.
Can I use Quivr with local models?
Yes, Quivr supports local models via Ollama by setting the appropriate environment variables.
What file types does Quivr support?
Quivr accepts any file type, including PDF, TXT, Markdown, and more. You can also add custom parsers.