mcp-ai-connector-dify

Model Context Protocol Gateway for Dify AI Services. This backend component allows abstract language models to interface with Dify AI's core chat generation capabilities via a standardized communication framework.

Core Capabilities

Connection bridge to Dify AI chat completion APIs
Specialized utility for venue location suggestions (named 'meshi-doko')
Mechanisms for maintaining dialogue history and session continuity
Support for chunked/streaming output transmission
Built entirely on TypeScript for robust typing

Deployment Instructions

Via Containerization (Docker)

# Produce the Docker image artifact
make docker

# Execute the service using Docker
docker run -i --rm mcp/dify https://your-dify-api-endpoint your-dify-api-key

Operational Use

Integration with Desktop Client (Example: Claude Desktop)

Incorporate the following configuration stanza into your claude_desktop_config.json:

{
  "mcpServers": {
    "dify": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-dify",
        "https://your-dify-api-endpoint",
        "your-dify-api-key"
      ]
    }
  }
}

Substitute placeholder values for your-dify-api-endpoint and your-dify-api-key with your valid credentials.

Integrated Utilities

meshi-doko (Venue Locator)

Utility designed for querying location-based information through the Dify AI interface:

Required Arguments: - LOCATION (Text): Geographical area for the search - BUDGET (Text): Financial constraints for the venue - query (Text): Primary prompt sent to the underlying model - conversation_id (Text, Optional): Identifier used for context threading

Development Lifecycle

# Set up initial project dependencies
make setup

# Compile source code
make build

# Apply code styling rules
make format

# Execute static analysis checks
make lint

Licensing

This software is distributed under the terms of the MIT License.

Security Considerations

This gateway handles sensitive Dify AI authentication tokens. Adherence to the following is crucial: - Secure storage of all access credentials is mandatory - Connection endpoints must utilize TLS/HTTPS exclusively - API secrets must never be committed to repository history