{"type":"mcp_server","name":"Deepseek-Thinking-Claude-3.5-Sonnet-CLINE-MCP","description":"This repository implements a Model Context Protocol (MCP) server, leveraging DeepSeek R1 for reasoning and Claude 3.5 Sonnet for response generation via OpenRouter. It manages conversations, optimizes parameters, and supports","category":"AI","language":"JavaScript","stars":114,"forks":20,"owner":"newideas99","github_url":"https://github.com/newideas99/Deepseek-Thinking-Claude-3.5-Sonnet-CLINE-MCP","homepage":null,"setup":"## Setup\n\n**Installing via Smithery:**\n\nTo install DeepSeek Thinking with Claude 3.5 Sonnet for Claude Desktop automatically via [Smithery](https://smithery.ai/server/@newideas99/Deepseek-Thinking-Claude-3.5-Sonnet-CLINE-MCP):\n\n```bash\nnpx -y @smithery/cli install @newideas99/Deepseek-Thinking-Claude-3.5-Sonnet-CLINE-MCP --client claude\n```\n\n**Manual Installation:**\n\n1. Clone the repository:\n```bash\ngit clone https://github.com/yourusername/Deepseek-Thinking-Claude-3.5-Sonnet-CLINE-MCP.git\ncd Deepseek-Thinking-Claude-3.5-Sonnet-CLINE-MCP\n```\n\n2. Install dependencies:\n```bash\nnpm install\n```\n\n3. Create a `.env` file with your OpenRouter API key:\n```env\n# Required: OpenRouter API key for both DeepSeek and Claude models\nOPENROUTER_API_KEY=your_openrouter_api_key_here\n\n# Optional: Model configuration (defaults shown below)\nDEEPSEEK_MODEL=deepseek/deepseek-r1  # DeepSeek model for reasoning\nCLAUDE_MODEL=anthropic/claude-3.5-sonnet:beta  # Claude model for responses\n```\n\n4. Build the server:\n```bash\nnpm run build\n```\n\nAdd to your Cline MCP settings (usually in `~/.vscode/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json`):\n\n```json\n{\n  \"mcpServers\": {\n    \"deepseek-claude\": {\n      \"command\": \"/path/to/node\",\n      \"args\": [\"/path/to/Deepseek-Thinking-Claude-3.5-Sonnet-CLINE-MCP/build/index.js\"],\n      \"env\": {\n        \"OPENROUTER_API_KEY\": \"your_key_here\"\n      },\n      \"disabled\": false,\n      \"autoApprove\": []\n    }\n  }\n}\n```","tools":"## Available Tools\n\n1.  Two-Stage Processing: Uses DeepSeek R1 for initial reasoning and Claude 3.5 Sonnet for final response through OpenRouter.\n2.  Smart Conversation Management: Detects active conversations, handles multiple concurrent conversations, filters out ended conversations, and supports context clearing.\n3.  Optimized Parameters: Uses model-specific context limits and recommended settings for temperature, top_p, and repetition_penalty.\n4.  `generate_response`: Main tool for generating responses with parameters for prompt, showing reasoning, clearing context, and including history.\n5.  `check_response_status`: Tool for checking the status of a response generation task using a task ID.\n6.  Response Polling: Uses a polling mechanism to handle long-running requests, providing task IDs and status updates.","faq":null,"created_at":"2025-01-25T17:16:46+00:00","updated_at":"2025-03-25T14:02:51+00:00","source_url":"https://model-context-protocol.com/servers/deepseek-claude-reasoning-retrieval-augmented-thinking","related_articles":[]}