> ## Documentation Index
> Fetch the complete documentation index at: https://docs.sequentum.com/llms.txt
> Use this file to discover all available pages before exploring further.

# MCP

> Wire Sequentum into LLM-driven data pipelines and analyst workflows through a single MCP server.

Sequentum exposes a [Model Context Protocol](https://modelcontextprotocol.io) server at `https://mcp.sequentum.com/mcp`. It's a typed wrapper over the [REST API](/api-reference/introduction) — same operations (list, run, schedule, monitor agents), surfaced as MCP tools that any MCP-aware LLM can call.

The primary use case is **automating Sequentum from inside an LLM-driven pipeline**. A Python service or n8n workflow calls Claude with the Sequentum MCP server attached; the model picks the right agent, runs it, and your pipeline gets structured records back — no separate REST integration. Interactive use from Claude Desktop is a useful secondary surface for analysts; dev-tool clients (Cursor, Claude Code) are fine for prototyping but aren't the production path.

<CardGroup cols={2}>
  <Card title="Connect" icon="plug" href="/mcp/connect">
    Wire Sequentum into a pipeline, automation platform, or chat client — hosted or self-hosted.
  </Card>

  <Card title="Tools" icon="toolbox" href="/mcp/tools">
    Every MCP tool the server exposes, mapped to its REST equivalent.
  </Card>
</CardGroup>

## Where to use MCP

| Pattern                                                                     | Use it for                                                                           |
| --------------------------------------------------------------------------- | ------------------------------------------------------------------------------------ |
| **Claude API + Sequentum MCP server**                                       | Production pipelines where an LLM orchestrates Sequentum runs. The killer use case.  |
| **Automation platforms with MCP support** (n8n, Zapier, AWS Step Functions) | No-code / low-code pipeline glue.                                                    |
| **Claude Desktop**                                                          | Interactive analyst workflows — ad-hoc data pulls, run inspection, schedule changes. |
| **Self-hosted (`npx -y sequentum-mcp`)**                                    | Air-gapped or on-prem environments that can't reach `mcp.sequentum.com`.             |

For straight code-driven integration without an LLM in the loop, the [REST API](/api-reference/introduction) is the simpler choice. MCP earns its keep when a model is already orchestrating.

[Wire it up →](/mcp/connect)