docs: Add community MCP server for Replicate

[deepfates]

Description

Add Replicate MCP server that lets language models search and run machine learning models through a simple tool-based interface. The server provides tools for browsing models, creating predictions, tracking their status, and handling generated images.

Server Details

• Server: Replicate
• Changes: New server implementation

Motivation and Context

This server makes Replicate's model catalog accessible to language models through MCP. It provides a structured way to:

• Search and browse available models
• Run models with text or structured input
• Track prediction status and results
• Handle generated images efficiently

How Has This Been Tested?

Tested with Claude Desktop client using common scenarios:

• Searching for text-to-image models
• Running predictions with various inputs
• Checking prediction status
• Viewing generated images
• Browsing model collections

Breaking Changes

None - this is a new server implementation.

Types of changes

• New feature (non-breaking change which adds functionality)

Checklist

• I have read the MCP Protocol Documentation
• My changes follows MCP security best practices
• I have updated the server's README accordingly
• I have tested this with an LLM client
• My code follows the repository's style guidelines
• New and existing tests pass locally
• I have added appropriate error handling
• I have documented all environment variables and configuration options

Additional context

The server requires a Replicate API token set as REPLICATE_API_TOKEN environment variable. All tools follow MCP best practices for error handling and input validation.