quivr
An opinionated RAG framework for integrating generative AI into applications, supporting any LLM, vector store, and file type.
What is quivr?
Quivr is an opinionated Retrieval-Augmented Generation (RAG) framework that simplifies adding generative AI capabilities to applications. It handles document ingestion, retrieval, and generation workflows, allowing developers to integrate AI assistants or knowledge bases without building RAG systems from scratch. The framework supports multiple LLMs, vector stores, and file formats for flexibility.
Developers and product teams building AI-powered applications like chatbots, knowledge assistants, or document analysis tools who want a pre-built, customizable RAG solution.
Quivr reduces development time by providing a production-ready RAG core with extensive customization options, avoiding vendor lock-in through support for any LLM or vector store. Its opinionated design ensures best practices while allowing integration into existing products.
Overview
An open-source engine that turns continuous content streams into search and monitoring. Durable ingestion, hybrid search, alerts, and plugins for formats, models and business rules.
Use Cases
Best For
- Adding AI-powered Q&A to existing applications
- Building custom knowledge bases from documents
- Creating chatbots with document retrieval capabilities
- Prototyping RAG applications quickly
- Integrating multiple LLMs into a single workflow
- Processing diverse file types (PDF, TXT, Markdown) for AI tasks
Not Ideal For
- Projects requiring standalone LLM interactions without document retrieval (e.g., simple chatbots with no knowledge base)
- Teams needing a fully managed, no-code AI platform with built-in user interfaces
- High-scale enterprise systems where fine-grained performance tuning and low-latency are critical
Pros & Cons
Pros
Provides a pre-built, fast RAG implementation that reduces development overhead, letting developers focus on product features rather than RAG complexities.
Supports multiple providers like OpenAI, Anthropic, and local models via Ollama, avoiding vendor lock-in and offering flexibility.
Handles various file types including PDF, TXT, and Markdown, with options for custom parsers to extend data ingestion.
Allows extensions like internet search and tools through YAML configuration, enabling tailored RAG applications beyond basic retrieval.
Cons
Requires YAML files and Python code for customization, which can be steep for developers unfamiliar with RAG concepts or workflow orchestration.
The project is actively evolving, as noted in the README, which might lead to breaking changes and incomplete features during updates.
Relies on third-party services for LLMs and rerankers (e.g., Cohere), introducing potential latency, cost, and reliability issues in production.
Frequently Asked Questions
Found a gem we're missing?
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.