Influencers Club MCP Server

Beta — This project is under active development. It works and can be tested, but expect changes before the stable release.

MCP server for the Influencers Club API — creator enrichment, discovery, audience analysis, content data, batch operations, and account management.

Prerequisites

• Python 3.10+ or Docker Desktop

• An Influencers Club API key

Quick Start (pip)

pip install influencers-club-mcp

Or install from source:

git clone https://github.com/Influencers-Club/influencers-club-mcp.git
cd influencers-club-mcp
pip install -e .

Then add to your Claude Desktop claude_desktop_config.json:

{
  "mcpServers": {
    "influencers-club": {
      "command": "influencers-club-mcp",
      "env": {
        "INFLUENCERS_CLUB_API_KEY": "your_api_key_here"
      }
    }
  }
}

Restart Claude Desktop — the server will appear as "influencers-club". That's it.

Where do files go? By default, exported CSVs and uploads are saved to exports/ and imports/ inside the cloned repo folder. No extra configuration needed.

To use a custom location, add these env vars:

"EXPORT_HOST_DIR": "/your/custom/exports",
"IMPORT_HOST_DIR": "/your/custom/imports",
"OUTPUT_DIR": "/your/custom/exports",
"IMPORTS_DIR": "/your/custom/imports"

Other Clients

Claude Code

Add to your project's .mcp.json or global ~/.claude/settings.json:

{
  "mcpServers": {
    "influencers-club": {
      "command": "influencers-club-mcp",
      "env": {
        "INFLUENCERS_CLUB_API_KEY": "your_api_key_here"
      }
    }
  }
}

VS Code / Cursor

Add to .vscode/mcp.json in your project:

{
  "servers": {
    "influencers-club": {
      "command": "influencers-club-mcp",
      "env": {
        "INFLUENCERS_CLUB_API_KEY": "your_api_key_here"
      }
    }
  }
}

Docker (Advanced)

Use Docker if you want filesystem isolation or prefer containerized deployments.

git clone https://github.com/Influencers-Club/influencers-club-mcp.git
cd influencers-club-mcp
docker build -t influencers-club-mcp .

Add to your Claude Desktop claude_desktop_config.json:

{
  "mcpServers": {
    "influencers-club": {
      "command": "docker",
      "args": [
        "run", "-i", "--rm",
        "-p", "127.0.0.1:8090:8090",
        "-e", "INFLUENCERS_CLUB_API_KEY=your_api_key_here",
        "-e", "UPLOAD_PORT=8090",
        "-e", "UPLOAD_BIND=0.0.0.0",
        "-v", "/path/to/exports:/exports",
        "-v", "/path/to/imports:/imports",
        "-e", "EXPORT_HOST_DIR=/path/to/exports",
        "-e", "IMPORT_HOST_DIR=/path/to/imports",
        "influencers-club-mcp"
      ]
    }
  }
}

Note: Replace /path/to/exports and /path/to/imports with actual paths on your machine. The path appears 4 times — update all of them.

Examples by OS:

• macOS/Linux: /Users/john/influencers-club-mcp/exports

• Windows: C:\\Users\\John\\Desktop\\influencers-club-mcp\\exports (use double backslashes)

After configuring, restart your client. The server will appear as "influencers-club".

Available Tools (29)

Creator Discovery

Enrichment

Content Data

Batch Enrichment

File Management

Discovery Reference Data

Account

Usage Examples

Find fitness influencers on Instagram:

"Find 5 Instagram creators with 50k-500k followers who post about fitness and have an engagement rate above 3%."

Enrich a creator profile by handle:

"Get me the full profile for @cristiano on Instagram, including audience demographics."

Batch enrich a CSV of emails:

"I have a CSV of 3,000 emails to batch enrich." — Claude will open the upload page for you.

Find similar creators for campaign scaling:

"I like the creator @MrBeast on YouTube. Find 10 similar creators with at least 100k followers."

Compare audience overlap:

"Compare the audience overlap between @nike, @adidas, and @puma on Instagram." — Claude will generate a Venn diagram.

Get a creator's recent posts:

"Show me the latest posts from @garyvee on TikTok with engagement metrics."

Find all connected social accounts:

"What other social media accounts does @cristiano have linked to their Instagram?"

Check your remaining API credits:

"How many credits do I have left?"

Batch Enrichment Modes

For large files (100+ entries), Claude opens a browser upload page at http://localhost:8090. Drag-and-drop your CSV — Claude auto-detects the upload and starts processing. For smaller batches (under 100), Claude enriches them in parallel automatically.

The upload page accepts multi-column CSVs and automatically extracts the correct column.

Supported Platforms

Environment Variables

Privacy Policy

This MCP server is a thin client between Claude and the Influencers Club API. The sections below describe what the server itself does with your data; the Influencers Club platform's full data practices are governed by the Influencers Club Privacy Policy.

Data we collect. None. The server does not collect, persist, or transmit personally identifiable information about the human using Claude. The only inputs it receives are the tool arguments Claude sends (e.g. a creator handle or a CSV path) and the bearer token used to authenticate against the Influencers Club API.

How data is used. Tool arguments are forwarded directly to the Influencers Club API to fulfill the requested operation (discovery, enrichment, batch processing, etc.). Responses are returned to Claude. The server performs no analytics, profiling, or model training on this data.

Storage. The server keeps no database. The only data written to disk are: (a) CSV exports/imports the user explicitly creates via the file tools, written to local directories the user controls (exports/, imports/, or paths set by EXPORT_HOST_DIR / IMPORT_HOST_DIR); (b) a small .ic_config.json file storing the chosen export path. No conversation content, API responses, or credentials are persisted.

Third-party sharing. The server communicates with exactly one third party: api-dashboard.influencers.club. No data is sent to any other endpoint, analytics service, or telemetry provider. The bearer token is transmitted only to the Influencers Club API over HTTPS and is redacted from any error messages returned to Claude.

Retention. The server retains nothing in-process beyond the lifetime of a single tool call, with the exception of the user-controlled CSV files described above, which the user can delete at any time. Bearer tokens are read from environment variables (or, in hosted/HTTP mode, from the authenticated request) and are never written to disk.

Contact. Privacy questions and data-deletion requests: gjorgji.p@influencers.club. Security disclosures may also be filed via GitHub Issues.

Support

• Email: gjorgji.p@influencers.club

• GitHub Issues: github.com/Influencers-Club/influencers-club-mcp/issues

License

MIT