Thoughtful Claude

🚀 Thoughtful Claude - DeepSeek R1 Inference Server

An MCP server that enhances Claude's inference capabilities by integrating DeepSeek R1's advanced inference engine. This server enables Claude to access DeepSeek's state-of-the-art inference model, developed through large-scale reinforcement learning.

🚀 Quick Start

Thoughtful Claude is an MCP server designed to enhance Claude's inference capabilities by integrating DeepSeek R1's advanced inference engine. It allows Claude to access a powerful inference model developed through large-scale reinforcement learning.

✨ Features

• Advanced Inference Integration

  • Leverage DeepSeek R1's inference engine.
  • Seamlessly integrate with Claude's thinking process.
  • Handle complex multi-step inference tasks.

• Enterprise-Grade Security

  • Support environment variables (.env file).
  • Securely handle API keys.
  • Do not expose any keys in responses.

• MCP Protocol Support

  • Comprehensive MCP server implementation.
  • Handle streaming responses.
  • Proper error management.

• Modern Python Architecture

  • Use async/await for efficient processing.
  • Proper exception handling.
  • Clean code organization.

📦 Installation

Prerequisites

• Python 3.12+
• uv package manager
• DeepSeek API key (obtain from platform.deepseek.com)

Quick Start

# Clone the repository
git clone https://github.com/martinbowling/thoughtful-claude.git
cd thoughtful-claude

# Install MCP and dependencies
pip install "mcp[cli]" httpx python-dotenv

# Create a .env file containing the API key
echo "DEEPSEEK_API_KEY=your_key_here" > .env

# Install the MCP server using environment variables
mcp install server.py -f .env

The mcp install command will:

• Register the server in Claude Desktop.
• Set environment variables from .env.
• Configure the server to use the correct Python interpreter.

You can verify the installation by checking the 🔨 tool icon in the Claude Desktop interface.

💻 Usage Examples

Starting the Server

• The server will automatically start when properly configured in Claude Desktop.

Basic Workflow

• Claude receives a query requiring inference.
• The query is sent to DeepSeek R1 for advanced inference.
• The inference is returned to Claude wrapped in the <ant_thinking> tag.
• Claude integrates the inference into its response.

Example Queries

• Mathematical comparison: "Is 9.9 greater than 9.11?"
• Logical puzzle: "If all A are B, and some B are C, what can we conclude?"
• Complex analysis: "Compare and contrast quantum computing and classical computing."

🔧 Technical Details

Inference Pipeline

• Query Processing

  • Accept a structured format of context and questions.
  • Combine inputs for comprehensive inference.

• DeepSeek R1 Integration

  • Model: deepseek-reasoner
  • Streaming: Enable real-time processing
  • Maximum tokens: 1 (optimized for inference extraction)
  • Output: Structured inference content

Error Handling

• API Errors

  • Gracefully wrap errors in the <ant_thinking> tag.
  • Provide clear error messages for debugging.
  • Properly propagate exceptions.

• Connection Issues

  • Set timeout handling (default 30 seconds).
  • Automatically clean up streams.
  • Manage resources.