Claude Managed Agents is Anthropic's hosted way to run agents: your agents, their runs and their history live on Anthropic's platform. omnara is a free, open-source project (3,000+ stars on GitHub, Apache-2.0) that does the same job on machines you control, with Claude or any other model. Its README calls it "the open-source managed agent platform" and "a self-hostable alternative to Claude Managed Agents and OpenAI's Agents API".
Here's what it does, how to start it, a first agent to try, and the catch.
What it does
- Models: Anthropic (Claude), OpenAI, OpenRouter, Amazon Bedrock, LiteLLM, Ollama, or any compatible endpoint, including self-hosted ones.
- Tools: built-in shell, file, web search and fetch, and ask-a-person tools, plus MCP servers and your own tools.
- Skills, memory stores and subagents you can inspect.
- Durable state: every step is saved, so an agent resumes after a crash or a restart.
- Schedules: cron triggers with time zones.
- Integrations: Slack, Discord and GitHub, plus streaming, webhooks and per-tool permissions.
- Machines: a small daemon (omnarad) connects outward, so you don't open any inbound ports.
- Your own database: agent history stays in your Postgres when you self-host.
Option 1: self-host it (free)
- Install Docker with Docker Compose.
- Run: git clone https://github.com/omnara-ai/omnara.git
- Run: cd omnara
- Run: docker compose -f compose.yaml --profile app up -d
- Open http://localhost:8000 and sign up. The local verification links show in: docker compose logs api
- A new instance starts empty: add a model provider (for Claude, your Anthropic API key), a model and a machine, and share them with your project.
Free, for your email
Get the full setup steps + the copy-paste setup prompt
The rest of this guide (4 more sections) opens right here. The setup prompt, one paste into Claude that sets it all up, comes by email only. Use your real address.
50,000+ follow where these guides come from. No spam, unsubscribe in one click.