Claude Desktop API Integration via MCP

# Claude Desktop API Integration via MCP This project develops a Machine Communication Protocol (MCP) Server to enable seamless integration between Claude Desktop and Claude API, allowing users to access advanced features such as system prompts and conversation management. ## Credits This project is based on [Claude_Desktop_API_USE_VIA_MCP](https://github.com/mlobo2012/Claude_Desktop_API_USE_VIA_MCP) by [mlobo2012](https://github.com/mlobo2012). Special thanks to the original author for developing this integration approach. ## Key Features - Direct connection to Claude API through MCP - Conversation history tracking and management - System prompt support - Seamless switching between Professional Plan and API usage - Easy configuration with Claude Desktop - Support for latest Claude models (Claude 3 Opus, Sonnet, Haiku) - Support for Claude 3.5 Sonnet - FastAPI service for external access ## When to Use ### Professional Plan (Default) - Regular conversations in Claude Desktop - Basic usage within plan limitations - No special configuration required ### API Token (Through this MCP server) - When you need a longer context window - When you want to use system prompts - When you need to bypass rate limits - When you need advanced conversation management ## Installation Steps ### 1. Clone the Repository ```bash # Using VS Code # 1. Press Cmd + Shift + P (macOS) or Ctrl + Shift + P (Windows) # 2. Type "Git: Clone" # 3. Paste: [Your URL] # Or using terminal git clone [Your URL] cd mcp-p-anthropic-server ``` ### 2. Install Dependencies ```bash pip install -r requirements.txt ``` ### 3. Configure Environment ```bash # Copy environment template cp .env.example .env # Edit .env and add your API key # ANTHROPIC_API_KEY=your_api_key_here ``` ### 4. Configure Claude Desktop #### macOS Go to `~/Library/Application Support/Claude/` ```bash # Using Finder # 1. Press Cmd + Shift + G # 2. Enter: ~/Library/Application Support/Claude/ ``` #### Windows Go to `%APPDATA%\Claude\` 1. Create or edit `claude_desktop_config.json` 2. Copy content from `config/claude_desktop_config.json` 3. Update paths and API key ## Usage Guide ### Basic Usage #### Normal Claude Desktop Usage - Just chat with Claude normally - Uses your Professional plan - No special commands needed #### API Usage ``` @claude-api Please answer using API: What is the capital of France? ``` ### Advanced Features #### Using System Prompts ``` @claude-api {"system_prompt": "You are an expert fitness coach"} Create a workout plan ``` #### Conversation Management ``` # Start a new conversation @claude-api {"conversation_id": "project1"} Let's talk about Python # Continue existing conversation @claude-api {"conversation_id": "project1"} Tell me more # View conversation history @claude-api get_conversation_history project1 # Clear conversation @claude-api clear_conversation project1 ``` #### View Available Models ``` @claude-api list_available_models ``` ## Cost Management - API calls will use your Anthropic API credits and may incur charges - Use Professional plan for general inquiries - Use `@claude-api` only when you need: - Longer context window - Custom system prompts - Bypass rate limits ## Available MCP Tools ### Claude API Tools This system includes 5 tools for working with the Claude API that you can invoke via Claude Desktop chat with the `@claude-api` command followed by the tool name and required parameters: #### 1. query_claude Allows you to send a query directly to the Claude API with advanced options ``` @claude-api query_claude prompt="Your question" [conversation_id="desired_id"] [system_prompt="Your system prompt"] [model="model_name"] [temperature=0.7] [max_tokens=4096] ``` Parameters: - `prompt`: The question or request to send (required) - `conversation_id`: Conversation ID to track the conversation (default: "default") - `system_prompt`: Optional system prompt to guide Claude's behavior (optional) - `model`: Claude model to use (default: "claude-3-sonnet-20240229") - `temperature`: Creativity level of the response (0.0-1.0, default: 0.7) - `max_tokens`: Maximum tokens to generate (default: 4096) #### 2. clear_conversation Clears the specified conversation history ``` @claude-api clear_conversation [conversation_id="default"] ``` Parameters: - `conversation_id`: Conversation ID to clear (default: "default") #### 3. get_conversation_history Retrieves conversation history for the specified ID ``` @claude-api get_conversation_history [conversation_id="default"] ``` Parameters: - `conversation_id`: Conversation ID to view history (default: "default") #### 4. list_conversations Lists all available conversation IDs ``` @claude-api list_conversations ``` No additional parameters required #### 5. list_available_models Lists all available Claude models ``` @claude-api list_available_models ``` No additional parameters required ### Advanced Usage Examples #### Using Custom Model with System Prompt ``` @claude-api query_claude prompt="Analyze this data and provide recommendations" model="claude-3-opus-20240229" system_prompt="You are an expert data analyst, focused on providing concise and actionable insights" ``` #### Tracking Complex Conversation ``` @claude-api query_claude prompt="Start planning a project" conversation_id="project_planning" system_prompt="You are an experienced project manager specializing in software development planning" @claude-api query_claude prompt="What are the main risks to watch out for?" conversation_id="project_planning" @claude-api query_claude prompt="Create an initial timeline" conversation_id="project_planning" ``` - `query_claude`: Creates direct API calls to Claude - Supports system prompts - Conversation tracking - `clear_conversation`: Resets conversation history - Manage multiple conversation threads - `get_conversation_history`: Retrieves conversation logs - Debug conversation flow - `list_conversations`: Shows all existing conversations - `list_available_models`: Shows all available Claude models ## Development The main server imports are in `src/server.py` and the API server is in `src/api_server.py` To extend functionality, you can add new tools using the `@mcp.tool()` decorator Example of adding a new tool: ```python @mcp.tool() async def custom_tool(param: str) -> str: """ Description of custom tool Args: param: Parameter description """ try: # Tool implementation return result except Exception as e: return f"Error: {str(e)}" ``` ## Troubleshooting ### API Key Issues - Check your API key in `.env` - Check Claude Desktop configuration paths - Ensure API key has proper permissions ### Connection Issues - Check if MCP server is running - Check Python environment - Check Claude Desktop logs ### Usage Issues - Check `@claude-api` syntax is correct - Check conversation IDs - Check system prompt format ## Support For issues and questions: - Open a new issue in the repository - Check existing discussions - Review troubleshooting guide