> For the complete documentation index, see [llms.txt](https://docs.sealsecurity.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.sealsecurity.io/reference/mcp-server/what-it-is.md).

# What the MCP server is

The Seal MCP server, Seal's Model Context Protocol endpoint for AI clients.

The Seal MCP server is a Model Context Protocol (MCP) endpoint hosted by Seal. MCP is an open protocol for connecting AI clients to tool servers.

Any MCP-aware client (an IDE assistant, an in-house agent, a Claude Desktop client, and so on) can connect to the endpoint and get tools spanning the same areas as the Seal UI: list discovered packages, create and delete Sealing Rules, trigger Generate fix, query the Dashboard, generate reports, and integrate with GitHub. You also get search over the Seal documentation on the same endpoint, so no second server is needed for that.

## When to use MCP versus the Public API

For AI-mediated calls, use the MCP server. For non-AI, one-off, or batch integrations, the [Public API](/reference/public-api.md) is the simpler fit.

## Permissions and tenant scoping

MCP tool calls use the calling user's role, the same one the Seal UI honors. Read-only tools need the `view` permission; mutations require the matching write permission. See [User roles](/getting-started/users-and-sso/user-roles.md) for the role-to-permission mapping.

Every call is scoped to the calling user's tenant. There is no cross-tenant access.

## Related

* [Tool catalog](/reference/mcp-server/tool-catalog.md)
* [Using the MCP server from your own client](/reference/mcp-server/using-from-your-client.md)