These are the docs for the Metabase master branch. Some features documented here may not yet be available in the current release. Check out the docs for the current stable version, Metabase v0.59.
AI agent resources for embedding
If you use an AI coding agent, you can give the agent Metabase-specific context to help with embedding setup, upgrades, and migrations.
Agent skills
We’ve developed some agent skills to give your AI agents step-by-step playbooks for specific embedding tasks.
Available skills
| Skill | Description |
|---|---|
| SDK version upgrade | Upgrade your modular embedding SDK, including changelog checks and breaking change handling. |
| Full app → modular embedding | Migrate from full app embedding to modular embedding. |
| Modular embedding → SDK (React) | Migrate from script-based modular embedding to the React SDK. |
| Static → guest embeds | Migrate from static (signed) embeds to guest embeds. |
| SSO for embeds | Set up SSO authentication for embedded Metabase. (In development.) |
Browse all skills on the agent skills repo.
llms.txt
Agents can read the docs you’re reading now (and many agents have already read our docs), but we also publish llms.txts so that you can give your AI agent:
If you’re on a specific version (e.g., v0.58), you can use versioned llms.txt files scoped to that version’s docs:
https://www.metabase.com/docs/v0.58/llms.txthttps://www.metabase.com/docs/v0.58/llms-embedding-full.txt
Agents are not magic
Always review and validate the changes made by agents. Check that your application builds, tests pass, and the embedding works as expected before committing anything.
Further reading
Read docs for other versions of Metabase.