Guides
Task-oriented how-tos for working with the platform.
The Platform pages explain how Levain works. These guides show you how to use it — one task at a time.
Build and run
- Agents — create, version, and manage the lifecycle of your agents.
- Recipes — author the agent's node graph and debug it in the draft sandbox.
- Runs — trigger agents, monitor progress, track cost, and cancel cleanly.
- Automation — fire runs on a cron or from external events.
- Artifacts — publish, list, and download files an agent produces during a run.
Give the agent context
- Knowledge Base — seed the wiki, shape its schema, and read maintained pages from your agents.
- Integrations — connect external services like GitHub and MCP servers to your workspace.
Configure your workspace
- Authentication — API keys, workspaces, and who has access to what.
- Bring Your Own Key — route your agents through your own model provider account.
If something is missing, the API reference covers every endpoint with schemas and examples, and support@levainlabs.com reaches the team directly.