AI product Open source

Omnara

Omnara is an open-source managed agent platform for running AI agents as a service on any compatible model and machine. It provides a REST API, TypeScript SDK, CLI, dashboard, and MCP server for launching agents, sending messages, streaming events, and handling approvals.

View repository Visit site Mentioned in 1 video ↓

Overview

The platform runs agent workers that build model context, call models, and execute tools, while PostgreSQL stores each agent's event log and durable state so workers can resume agents after crashes, restarts, or machine disconnections. Agents can use built-in tools, skills, MCP servers, custom tools, persistent memory, subagents, permissions, schedules, webhooks, and integrations with Slack, Discord, and GitHub. Machines connect through the Omnara daemon and may be local laptops, virtual machines, containers, or managed sandboxes; agents can run across multiple machines without inbound ports.

Omnara Cloud is available at app.omnara.com, and the platform can be self-hosted with Docker Compose under the Apache License 2.0. Self-hosted deployments keep agent history in the operator's PostgreSQL database and can use any supported model provider or compatible self-hosted endpoint.

What Omnara is used for

1 use taken from transcripts — each links to the moment in the video.

  • A platform that runs AI agents as a service across local machines and cloud sandboxes, allowing job launching via API and live tool call monitoring.

Videos mentioning Omnara

1 in the library.