Quivr
Build opinionated RAG applications quickly using python and pluggable LLMs.
Integrate custom Retrieval-Augmented Generation workflows into your existing applications.
Quivr-core is a Python library designed to help you construct and configure RAG pipelines with minimal code. It suits developers looking to embed document querying and AI assistants into their applications.
What it is
The library lets you create a brain from PDF, TXT, or Markdown files using just a few lines of Python. It supports customizable retrieval strategies, YAML-based workflow definitions, and integration with the Megaparse ingestion tool.
Who it's for
It suits developers and engineers building custom AI products who want to bypass low-level RAG boilerplate. It does not suit non-technical users looking for a ready-to-use desktop chat application.
Key features
Opinionated RAG Framework
Builds fast, efficient RAG pipelines with minimal code using quivr-core.
Flexible LLM Support
Connects with OpenAI, Anthropic, Mistral, or local models via Ollama.
Any File Ingestion
Processes PDFs, text documents, and markdown files out of the box.
YAML Workflows
Defines retrieval strategies, history limits, and rerankers through configuration files.
Up and running in 3 steps
Our honest verdict
What's great
- Sets up a functional RAG pipeline in under ten lines of Python code.
- Supports multiple major LLM providers as well as local execution.
- Allows fine-grained control over retrieval configs and rerankers via YAML.
- Integrates directly with Megaparse for advanced document ingestion.
Where it falls short
- Requires writing Python code and handling environment variables to function.
- Depends on external commercial API keys for LLMs unless running local models.
- Documentation and setup assume familiarity with developer tooling.
Pricing & license
The software is free under its license, but you will incur costs for whichever LLM APIs or reranker services you choose to connect to it.
Open remote AI-training roles
880 open remote AI-training roles, updated November 15, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
What LLMs does Quivr support?
It supports APIs from OpenAI, Anthropic, and Mistral, alongside local models via Ollama.
Can I customize the retrieval workflow?
Yes, you can define custom retrieval strategies, history limits, and rerankers using a YAML configuration file.
