kotaemon
An open-source RAG-based tool for chatting with your documents.
Chat with your documents using local or API-based LLMs with advanced citations.
Kotaemon provides a clean, customizable web interface for retrieval-augmented generation (RAG) document question-answering. It suits developers building custom pipelines and end users needing a ready-to-use document chat application.
What it is
The application offers a Gradio-based web UI supporting multi-user logins, file organization in public or private collections, and hybrid retrievers combining full-text and vector search with re-ranking. It supports multi-modal parsing for figures and tables, detailed citations with an in-browser PDF viewer, question decomposition, and advanced agent reasoning methods like ReAct and ReWOO.
Who it's for
It suits developers looking for a modular RAG framework and end users wanting an interface to query internal documents. It is not suitable for users looking for zero-setup cloud SaaS without managing local infrastructure or API keys.
Key features
Hybrid RAG Pipeline
Combines full-text and vector search with re-ranking to maximize retrieval quality.
Advanced Citations
Provides detailed citations and highlights directly within an in-browser PDF viewer.
Multi-modal QA Support
Performs question answering on documents containing figures and tables.
Complex Reasoning
Supports question decomposition and agent-based reasoning using ReAct and ReWOO.
Up and running in 3 steps
Our honest verdict
What's great
- Supports both local LLMs via Ollama and popular API providers like OpenAI and Azure.
- Includes multi-user login and private or public file collection management.
- Built on Gradio for straightforward UI customization and theme extension.
- Features GraphRAG and NanoGraphRAG indexing pipeline examples.
Where it falls short
- Requires Python 3.10+ and environment file configuration for local deployments.
- Requires installing Unstructured dependencies for processing non-PDF file formats.
- In-browser PDF_JS viewer requires manual asset downloading and extraction.
Pricing & license
Free under the Apache-2.0 license. The only cost is running it yourself on your own hardware or paying for LLM API providers if you choose not to run local models.
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
Use it, or have us set it up
FAQ
What file types can Kotaemon process?
It handles PDF, HTML, MHTML, and XLSX documents out of the box, with additional formats supported by installing Unstructured.
Can I run Kotaemon locally without external API keys?
Yes, you can use local LLMs via Ollama or llama-cpp-python, or use the Docker image bundled with Ollama.
