Connects to over 50 applications to provide intelligent search and AI assistance across company information.

Onyx is an open-source web app and server that provides a comprehensive application layer for Large Language Models. It functions as a search and AI assistance platform that connects to over 50 indexing connectors and Model Context Protocol (MCP) sources to aggregate company information into a single interface. The software is deployed as a server and accessed via a web application, allowing organizations to maintain control over their data and AI infrastructure.
Users can deploy the platform in two distinct modes depending on their resource availability and needs. The Lite version serves as a lightweight chat interface requiring less than 1GB of memory, making it suitable for quick testing or teams only requiring basic agent functionalities. The Standard version provides the full feature set, including vector and keyword indexing for Retrieval-Augmented Generation (RAG), background containers for job queues, and AI model inference servers. This version also integrates with Redis for in-memory caching and MinIO for blob storage to optimize performance for large scale use.
The system architecture supports Docker, Kubernetes, and Helm deployments, with specific guides provided for major cloud providers. It is designed for a wide range of users, from individual developers to large global organizations requiring audit logs, query history, and analytics broken down by team or model. The platform allows for the implementation of custom code to handle PII removal, reject sensitive queries, or perform specialized data analysis.
Onyx is available in a Community Edition under the MIT license and a separate Enterprise Edition for larger organizations requiring advanced administrative controls.
A self-hosted interface for interacting with local and cloud LLMs with built-in RAG and tool support.
A cross platform interface for interacting with multiple large language models via API or local deployment.
Join our newsletter to get shiny new open source software delivered to your inbox. Unsubscribe anytime.