Skip to main content
Glama
HarshJ23

DeepSeek-Claude MCP Server

by HarshJ23
README.md
# DeepSeek-Claude MCP Server
[![smithery badge](https://smithery.ai/badge/@HarshJ23/deepseek-claude-MCP-server)](https://smithery.ai/server/@HarshJ23/deepseek-claude-MCP-server)

**Enhance Claude's reasoning capabilities** with the integration of DeepSeek R1's advanced reasoning engine. This server enables Claude to tackle complex reasoning tasks by leveraging the reasoning capabilites of deepseek r1 model.

<a href="https://glama.ai/mcp/servers/w3q7v1srsw"><img width="380" height="200" src="https://glama.ai/mcp/servers/w3q7v1srsw/badge" alt="DeepSeek-Claude Server MCP server" /></a>

---

## πŸš€ Features

### **Advanced Reasoning Capabilities**
- Seamlessly integrates DeepSeek R1's reasoning with Claude.
- Supports intricate multi-step reasoning tasks.
- Designed for precision and efficiency in generating thoughtful responses.






---

## Complete Setup guide

### Installing via Smithery

To install DeepSeek-Claude for Claude Desktop automatically via [Smithery](https://smithery.ai/server/@HarshJ23/deepseek-claude-MCP-server):

```bash
npx -y @smithery/cli install @HarshJ23/deepseek-claude-MCP-server --client claude
```

### Prerequisites
- Python 3.12 or higher
- `uv` package manager
- DeepSeek API key (Sign up at [DeepSeek Platform](https://platform.deepseek.com))



1. **Clone the Repository**
   ```bash
   git clone https://github.com/harshj23/deepseek-claude-MCP-server.git
   cd deepseek-claude-MCP-server
   ```

2. **Ensure UV is Set Up**
   - **Windows**: Run the following in PowerShell:
     ```powershell
     powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
     ```
   - **Mac**: Run the following:
     ```bash
     curl -LsSf https://astral.sh/uv/install.sh | sh
     ```

3. **Create Virtual Environment**
   ```bash
   uv venv
   source .venv/bin/activate
   ```

4. **Install Dependencies**
   ```bash
   uv add "mcp[cli]" httpx
   ```

5. **Set Up API Key**
   ```bash
   Obtain your api key from here : https://platform.deepseek.com/api_keys
   ```

6. **Configure MCP Server**
   Edit the `claude_desktop_config.json` file to include the following configuration:
   ![claude_desktop_config.json file](public/images/edit_config.png)

   ```json
   {
       "mcpServers": {
           "deepseek-claude": {
               "command": "uv",
               "args": [
                   "--directory",
                   "C:\\ABSOLUTE\\PATH\\TO\\PARENT\\FOLDER\\deepseek-claude",
                   "run",
                   "server.py"
               ]
           }
       }
   }
   ```

7. **Run the Server**
   ```bash
   uv run server.py
   ```

8. **Test Setup**
   - ##### Restart Claude Desktop.
   - Verify the tools icon is visible in the interface.
   ![tool visible](public/images/tool1.png)
![tool verify](public/images/tool2.png)

   - If the server isn’t visible, consult the [troubleshooting guide](https://modelcontextprotocol.io/quickstart/server#troubleshooting).

---

## πŸ›  Usage

### Starting the Server
The server automatically starts when used with Claude Desktop. Ensure Claude Desktop is configured to detect the MCP server.

### Example Workflow
1. Claude receives a query requiring advanced reasoning.
2. The query is forwarded to DeepSeek R1 for processing.
3. DeepSeek R1 returns structured reasoning wrapped in `<ant_thinking>` tags.
4. Claude integrates the reasoning into its final response.

---


## πŸ“„ License

This project is licensed under the MIT License. See the [LICENSE](LICENSE) file for details.

---

TDQS

B3.4/5.0

Scored across 1 tool

Disambiguation5/5

With only one tool named 'reason', there is no possibility of confusion or overlap with other tools. The tool has a single, clearly defined purpose: processing queries through DeepSeek's R1 reasoning engine for Claude integration.

Naming Consistency5/5

A single tool inherently demonstrates perfect naming consistency. The tool name 'reason' follows a clear verb-based pattern appropriate for its function, and there are no other tools to create inconsistency.

Tool Count2/5

A single tool server is generally too minimal for most practical applications. While the tool itself performs a specific reasoning task, the server lacks complementary tools for broader reasoning workflows, making it feel incomplete as a standalone server.

Completeness2/5

The server is severely incomplete for a reasoning engine interface. It provides only query processing without any supporting tools for configuration, history management, different reasoning modes, or result validation. This creates significant gaps that will limit agent effectiveness.

Maintenance

ActivityInactive
ResponsivenessNo issues