Open-source AI assistant that connects to 15+ work tools. Search across your organization, get AI-powered answers, and automate tasks. Self-hosted on your infrastructure.
Omni is an AI agent platform powered by deep search across all your work tools. Ask questions, run code, automate tasks, and always get answers grounded in your organization's real data.
Ask anything about your work and get instant answers with source citations across all connected apps.
Use Anthropic, OpenAI, Gemini, Azure, or self-host open-weight models with vLLM. Your choice, your infrastructure.
The AI agent can execute Python and bash in isolated, sandboxed containers to analyze data and automate tasks.
Omni inherits permissions from every connected source system. Users only ever see data they already have access to, enforced in real time across every search and agent interaction.
Rich search interface with filters by source, date, type, and more. Find exactly what you need, fast.
Model Context Protocol for seamless tool integration with any MCP-compatible AI client.
Build custom connectors in Python or TypeScript to index any internal or external data source.
Docker Compose for development. Terraform modules for production deployments on AWS or GCP.
ParadeDB with BM25 full-text search and pgvector semantic search. No Elasticsearch needed.
Apache 2.0 licensed. Every component is auditable, from data ingestion to AI processing.
Omni plugs into 15+ tools your team already uses, powering search, AI agents, and automation across your organization. Need something custom? Build it with our SDK.
Don't want to self-host? Sign up for early access to Omni Cloud. Fully managed, so you can focus on your work.