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

# MCP Server

> Connect Genseo to MCP clients with project-bound tools.

The Genseo MCP server lets AI tools use Genseo through structured tools instead of raw HTTP calls. It uses the same Public API and the same project-bound API key rules.

An MCP client can only access the project attached to the key you provide.

## Global install

Install the CLI package globally:

```bash theme={null}
npm install -g @genseo/cli
```

Then configure your MCP client:

```json theme={null}
{
  "mcpServers": {
    "genseo": {
      "command": "genseo-mcp",
      "env": {
        "GENSEO_API_BASE": "https://api.genseo.co/v1",
        "GENSEO_API_KEY": "gs_live_..."
      }
    }
  }
}
```

## Without global install

Use `npx` if you do not want a global package:

```json theme={null}
{
  "mcpServers": {
    "genseo": {
      "command": "npx",
      "args": ["-y", "--package", "@genseo/cli", "genseo-mcp"],
      "env": {
        "GENSEO_API_BASE": "https://api.genseo.co/v1",
        "GENSEO_API_KEY": "gs_live_..."
      }
    }
  }
}
```

## Available tools

* `genseo_me`
* `genseo_project_get`
* `genseo_keywords_list`
* `genseo_keywords_create`
* `genseo_keywords_generate`
* `genseo_keyword_metrics`
* `genseo_posts_list`
* `genseo_posts_create_draft`
* `genseo_posts_get`
* `genseo_posts_update`
* `genseo_posts_generate`
* `genseo_posts_publish`
* `genseo_integrations_list`
* `genseo_integration_fields`
* `genseo_integration_connect_url`
* `genseo_integration_mapping_save`
* `genseo_webhooks_list`
* `genseo_webhooks_create`
* `genseo_webhooks_delete`

## Agent behavior

MCP clients should call `genseo_me` first, use only the returned project, create drafts before publishing, and ask before publishing unless the user has explicitly enabled autonomous publishing.