Bauplan is driven entirely by code, which makes it a natural fit for AI agents. The whole lakehouse is programmable through a [CLI](/reference/cli) and a [Python SDK](/reference/bauplan): there is no UI step an agent has to click through.

[Git-for-data branching](/concepts/git-for-data/) is what makes it safe to hand that control to an agent. The agent works on an isolated data branch, validates its changes with [expectations](/concepts/expectations) and the Write-Audit-Publish pattern, and merges into the `main` branch only when the checks pass. Nothing reaches production data until it has been reviewed, and every change is a commit you can inspect or revert.

There are three complementary ways to use agents with Bauplan. They are not mutually exclusive: an assistant can use the MCP server for live access, run Skills for structured workflows, and read the documentation as context, all in the same session.

,
    description: 'Reusable recipes for the main workflows, including pipelines, ingestion, exploration, and debugging.'
  },
  {
    type: 'link',
    href: '/agents/mcp',
    label: 'MCP server',
    icon: ,
    description: 'Give an assistant live access to your lakehouse through tool calls.'
  },
  {
    type: 'link',
    href: '/agents/context',
    label: 'Context for agents',
    icon: ,
    description: 'Feed AGENTS.md, llms.txt, and the Markdown docs to any LLM or IDE assistant.'
  }
]} />

## Learn more

The [Bauplan blog](https://www.bauplanlabs.com/blog) covers running AI and agents safely on your data, alongside technical deep-dives on the platform. The team also publishes peer-reviewed research on the data systems behind Bauplan and their application to AI and agentic workflows. Find our latest results on [Google Scholar](https://scholar.google.com/scholar?hl=en&as_sdt=0%2C5&q=bauplan+lakehouse&btnG=).