Onyx

Description

Onyx is an open-source AI platform that unifies search and chat across a company's internal tools and documents. It connects to sources like Slack, Confluence, Google Drive, and GitHub, then lets any LLM answer questions grounded in that private knowledge instead of the public internet.

Features

• 40+ Connectors: Indexes content from Slack, Confluence, Notion, Google Drive, GitHub, and more.
• Any LLM: Works with OpenAI, Anthropic, or fully local open-weight models.
• Permission-Aware Search: Respects source-system permissions so users only see what they're allowed to.
• Chat & Search Modes: Offers both a conversational assistant and a traditional search interface.
• Continuous Sync: Keeps indexed content up to date as source documents change.
• Self-Hosted Deployment: Runs fully on your own infrastructure for data-sensitive environments.

Technology Stack

Backend built in Python with FastAPI; frontend built with Next.js and TypeScript. Uses Postgres for metadata, Vespa for search indexing, and supports any LLM via a pluggable model layer.

Requirements

Linux server with Docker Compose or Kubernetes; at least 8 GB RAM for a small deployment; Postgres database; more resources needed as connector count and document volume grow.