Replicate Mcp Server

Replicate Mcp Server

What is Replicate Mcp Server about?

Provides a Model Context Protocol (MCP) server implementation that lets users interact with Replicate's AI models. The server offers tool‑based commands for searching models, creating and monitoring predictions, and handling generated images, all usable from MCP‑compatible clients such as Claude Desktop, Cursor, Cline, or Continue.

How to use Replicate Mcp Server?

Installation

npm install -g mcp-replicate   # or use npx for a one‑off run

Or run directly without installing:

npx -y mcp-replicate

Configuration

1. Obtain a Replicate API token from the Replicate account page.
2. Add the token to the MCP server configuration (recommended via Claude Desktop) or export it as an environment variable:

{
  "mcpServers": {
    "replicate": {
      "command": "mcp-replicate",
      "env": { "REPLICATE_API_TOKEN": "your_token_here" }
    }
  }
}

or export REPLICATE_API_TOKEN=your_token_here before launching the server. 3. Start the server (npx -y mcp-replicate) and ensure your MCP client points to it.

Key Features

• Model Tools: search_models, list_models, get_model, list_collections, get_collection.
• Prediction Tools: create_prediction, create_and_poll_prediction, get_prediction, cancel_prediction, list_predictions.
• Image Tools: view_image, clear_image_cache, get_image_cache_stats.
• Semantic search for models, version details, and collection browsing.
• Real‑time prediction status tracking and cancellation.
• Browser‑based image preview and cache management for performance.

Use Cases

• Integrate Replicate model execution into Claude Desktop workflows, enabling AI‑assisted chat interactions.
• Rapid prototyping of AI predictions from the command line without writing custom API calls.
• Centralized management of generated images and cache handling in development environments.
• Building custom MCP clients that leverage the same tool set for diverse IDEs or automation scripts.

FAQ

Q: The server is running but tools don't appear in my client. A: Verify the MCP server settings in your client, ensure the REPLICATE_API_TOKEN is correctly set, and restart both the server and the client.

Q: Where should I set the Replicate API token? A: Prefer adding it to the client’s configuration file (as shown above). Alternatively, export it as an environment variable before launching the server.

Q: What Node.js version is required? A: Node.js >= 18.0.0.

Q: Can I run the server without installing it globally? A: Yes, use npx -y mcp-replicate for a one‑time execution.

Q: How do I view generated images? A: Use the view_image tool, which opens the image URL in the default web browser.

Replicate MCP Server

A Model Context Protocol server implementation for Replicate. Run Replicate models through a simple tool-based interface.

Quickstart

1. Install the server:

npm install -g mcp-replicate

2. Get your Replicate API token:

   • Go to Replicate API tokens page
   • Create a new token if you don't have one
   • Copy the token for the next step

3. Configure Claude Desktop:

   • Open Claude Desktop Settings (⌘,)
   • Select the "Developer" section in the sidebar
   • Click "Edit Config" to open the configuration file
   • Add the following configuration, replacing your_token_here with your actual Replicate API token:

{
  "mcpServers": {
    "replicate": {
      "command": "mcp-replicate",
      "env": {
        "REPLICATE_API_TOKEN": "your_token_here"
      }
    }
  }
}

4. Start Claude Desktop. You should see a 🔨 hammer icon in the bottom right corner of new chat windows, indicating the tools are available.

(You can also use any other MCP client, such as Cursor, Cline, or Continue.)

Alternative Installation Methods

Install from source

git clone https://github.com/deepfates/mcp-replicate
cd mcp-replicate
npm install
npm run build
npm start

Run with npx

npx mcp-replicate

Features

Models

• Search models using semantic search
• Browse models and collections
• Get detailed model information and versions

Predictions

• Create predictions with text or structured input
• Track prediction status
• Cancel running predictions
• List your recent predictions

Image Handling

• View generated images in your browser
• Manage image cache for better performance

Configuration

The server needs a Replicate API token to work. You can get one at Replicate.

There are two ways to provide the token:

1. In Claude Desktop Config (Recommended)

Add it to your Claude Desktop configuration as shown in the Quickstart section:

{
  "mcpServers": {
    "replicate": {
      "command": "mcp-replicate",
      "env": {
        "REPLICATE_API_TOKEN": "your_token_here"
      }
    }
  }
}

2. As Environment Variable

Alternatively, you can set it as an environment variable if you're using another MCP client:

export REPLICATE_API_TOKEN=your_token_here

Available Tools

Model Tools

• search_models: Find models using semantic search
• list_models: Browse available models
• get_model: Get details about a specific model
• list_collections: Browse model collections
• get_collection: Get details about a specific collection

Prediction Tools

• create_prediction: Run a model with your inputs
• create_and_poll_prediction: Run a model with your inputs and wait until it's completed
• get_prediction: Check a prediction's status
• cancel_prediction: Stop a running prediction
• list_predictions: See your recent predictions

Image Tools

• view_image: Open an image in your browser
• clear_image_cache: Clean up cached images
• get_image_cache_stats: Check cache usage

Troubleshooting

Server is running but tools aren't showing up

1. Check that Claude Desktop is properly configured with the MCP server settings
2. Ensure your Replicate API token is set correctly
3. Try restarting both the server and Claude Desktop
4. Check the server logs for any error messages

Tools are visible but not working

1. Verify your Replicate API token is valid
2. Check your internet connection
3. Look for any error messages in the server output

Development

1. Install dependencies:

npm install

2. Start development server (with auto-reload):

npm run dev

3. Check code style:

npm run lint

4. Format code:

npm run format

Requirements

• Node.js >= 18.0.0
• TypeScript >= 5.0.0
• Claude Desktop for using the tools

License

MIT