Omnara helps developers build AI agents without building the infrastructure to run them. Its open-source API handles agent execution and state, allowing agents to resume after interruptions. Choose the models and tools your agents use, run them on your own machines or managed sandboxes, and connect them to your application through the API. You can use Omnara Cloud or self-host the platform.
Key Features
- Agent API: Launch agents, send them instructions, and follow their work from your application.
- State and recovery: Keep conversation history and agent state so work can resume after a crash or disconnect.
- Model choice: Use supported model providers, bring your own API keys, or connect a compatible endpoint.
- Machine choice: Connect your own computer or server, or use a managed sandbox.
- Tools and integrations: Add custom tools, reusable skills, and MCP integrations.
- Hosting options: Use Omnara Cloud or self-host in your own environment.
Benefits
- Build agent features without creating your own execution and state management system.
- Use the models, tools, and machines that fit your application.
- Follow agent activity and require approval before selected actions.
- Self-host when you need more control over your infrastructure and agent data.
