Model Context Protocol (MCP) setup guide
Connect your MyContentLab marketing intelligence directly to Claude Desktop, Claude.ai Web, Claude Code CLI, or Cursor / Antigravity IDE using the open Model Context Protocol (MCP).
Overview
The MyContentLab MCP server enables AI assistants to query real-time workspace data — client accounts, multi-channel performance metrics, report summaries, brand listening mentions, market trends, CRM pipeline totals, and Content Lab ideas.
NOTE
Access is strictly read-only: AI assistants can look up and analyze data, but can never create, modify, or delete anything in your account.
Method 1: Claude Desktop and Claude.ai Web (OAuth 2.1 — recommended)
The fastest and most secure method uses the official OAuth 2.1 Remote Connector flow. No config files or static token copying required.
Step 1: Open connector settings
- In Claude Desktop or Claude.ai Web, go to: Settings → Connectors → Add custom connector (or Add Tool).
- Enter the connector configuration:
- Server URL:
https://app.mycontentlab.ai/functions/mcp-server - Authentication:
Always required(orRequired when the server asks) - OAuth client:
No client ID — register one automatically(uses RFC 7591 Dynamic Client Registration)
- Server URL:
Step 2: Authorize MyContentLab
- Click Connect / Add Connector.
- A browser window opens displaying the MyContentLab Consent Screen.
- Log in with your MyContentLab account (owner or manager role required).
- Review the requested permissions and click Allow Access.
Claude will automatically exchange tokens and register the tools. You can now mention @mycontentlab or ask questions directly in your chat.
Method 2: Claude Code CLI (terminal)
For automated or terminal workflows using Claude Code:
Step 1: Generate a Personal Access Token (PAT)
- Go to Settings → API & MCP Access in MyContentLab.
- Click New token, select the desired scopes and expiration (30, 90, or 365 days).
- Copy your token starting with
mcl_pat_....
Step 2: Register the server
Run the following command in your terminal (replace YOUR_PAT_TOKEN):
claude mcp add --transport http mycontentlab \
https://app.mycontentlab.ai/functions/mcp-server \
--header "Authorization: Bearer YOUR_PAT_TOKEN"Method 3: Claude Desktop via config file (PAT bridge)
If you prefer using static PAT credentials in Claude Desktop via mcp-remote:
- Open your configuration file:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
- macOS:
- Add the following server configuration:
{
"mcpServers": {
"mycontentlab": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"https://app.mycontentlab.ai/functions/mcp-server",
"--header",
"Authorization:${AUTH_HEADER}"
],
"env": {
"AUTH_HEADER": "Bearer YOUR_PAT_TOKEN"
}
}
}
}- Restart Claude Desktop completely.
Available MCP tools
| Tool name | Required scope | Description |
|---|---|---|
list_clients | read:clients | Lists all active clients, company names, industry, and websites. |
get_client_metrics | read:metrics | Retrieves monthly performance metrics across connected ad and social platforms. |
get_report_summary | read:metrics | Returns executive AI summaries, key findings, and action items from reports. |
list_mentions | read:mentions | Returns brand mentions and sentiment analysis from Social Listening. |
get_trends | read:trends | Queries market trends and opportunities relevant to client industry tags. |
get_crm_summary | read:crm | Aggregates deal stages, total pipeline value, and conversion metrics. |
list_content_ideas | read:contentlab | Returns AI-generated content ideas, hooks, angles, and opportunity scores. |
Sample prompts
- "List my active clients from MyContentLab."
- "Show last month's performance metrics for [Client Name] and summarize the top performing platform."
- "What are the latest social listening mentions for [Client Name] with negative sentiment?"
- "Give me the highest-scoring content ideas for [Client Name] from Content Lab."
- "Summarize the current CRM open pipeline and deal count for [Client Name]."

