DeepCo

DeepCo is a cross-platform chat client that supports configuring MCP servers to talk to various LLM providers. This MCP server acts as a bridge between DeepCo and your chosen language model backends (such as OpenAI, Grok, Google Gemini, DeepSeek, and other OpenAI-compatible APIs) as well as local/private models via LM Studio/Ollama. In DeepCo, open the MCP/Server area in the app and add a new server entry named 'deepco'. Configure the endpoint (by default the local endpoint the server runs on) and provide any required API keys or access tokens for providers you plan to use. The server exposes standard MCP endpoints for chat, prompt management, and character adaptation (via SillyTavern assets) so you can route conversations to the appropriate LLM backend, manage prompts, and use predefined characters during chats.

Prerequisites:

• macOS: brew install uv
• macOS: brew install node (if you plan to use Node-based tooling alongside)
• Windows: winget install --id=astral-sh.uv -e and install Node.js LTS
• Linux: ensure uv and Node.js are available via your distro's package manager

Install and run DeepCo MCP server:

1. Install prerequisites (uv and Node.js) according to your OS:
   • macOS: brew install uv && brew install node
   • Windows: follow MCP ENV instructions to install uv and Node.js
   • Linux: use your distro's package manager to install uv and Node.js
2. Install the MCP server package (example using uvx): uvx install succlz123/deepco-mcp-server
3. Run the MCP server: uvx run succlz123/deepco-mcp-server
4. In the DeepCo desktop app, go to the MCP settings and add a server named 'deepco' pointing to the local endpoint (default port shown by the server). Provide any necessary environment variables (API keys, endpoints) in the server configuration.
5. If needed, configure provider keys (OPENAI_API_KEY, GROK_API_KEY, GEMINI_API_KEY, etc.) as environment variables for the MCP server.

Notes:

• The actual package name and run command may vary; follow the exact commands shown by your package manager when installing the 'deepco-mcp-server' package.

Environment variables you may need:

• OPENAI_API_KEY (for OpenAI/OpenAI-compatible providers)
• GROK_API_KEY
• GEMINI_API_KEY
• OPENROUTER_API_KEY or other provider keys as applicable
• Any provider-specific BASE_URL or MODEL_NAME configurations Common issues:
• Server failing to start due to missing API keys—provide keys in the MCP server config or environment before starting.
• Port conflicts—ensure the MCP server port is not in use by another service and that the DeepCo app points to the correct local endpoint.
• Firewall or network restrictions may block outbound requests to provider APIs.