LLM Bridge

LLM Bridge MCP provides a unified interface to access multiple LLM providers through the MCP protocol. You can connect your client to different models, switch providers easily, and track usage all from a single server implementation.

How to use

You use an MCP client to connect to the LLM Bridge MCP server. The server exposes a stdio entry point that you run locally, as shown in the configuration example. You can supply different API keys for supported providers and adjust model parameters such as temperature and max tokens for each request.

Key capabilities you can leverage include sending prompts to various models, selecting a specific model name, and controlling model behavior with a system prompt. The server provides a single function interface you can call from your client to obtain a response from the chosen provider.

How to install

Prerequisites you need before installation: Node.js and npm present on your system, and the uv runtime installed for local MCP execution.

Option A: Install via Smithery for automatic setup with Claude Desktop support.

npx -y @smithery/cli install @sjquant/llm-bridge-mcp --client claude

Option B: Manual installation steps.

git clone https://github.com/yourusername/llm-bridge-mcp.git
cd llm-bridge-mcp

Install uv if it is not already installed on your system. Follow the platforms below.

# On macOS
brew install uv

# On Linux
curl -LsSf https://astral.sh/uv/install.sh | sh

# On Windows
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

Environment and configuration

Create a configuration file to supply your API keys for the supported providers. This example shows the keys you will typically need to set as environment variables.

OPENAI_API_KEY=your_openai_api_key
ANTHROPIC_API_KEY=your_anthropic_api_key
GOOGLE_API_KEY=your_google_api_key
DEEPSEEK_API_KEY=your_deepseek_api_key

Usage with Claude Desktop or Cursor

Add a server entry to your Claude Desktop or Cursor MCP configuration so you can run the MCP server locally and connect to your preferred providers.

"mcpServers": {
  "llm-bridge": {
    "command": "uvx",
    "args": [
      "llm-bridge-mcp"
    ],
    "env": {
      "OPENAI_API_KEY": "your_openai_api_key",
      "ANTHROPIC_API_KEY": "your_anthropic_api_key",
      "GOOGLE_API_KEY": "your_google_api_key",
      "DEEPSEEK_API_KEY": "your_deepseek_api_key"
    }
  }
}

Troubleshooting

Common issues you might run into include missing the uvx executable in your PATH. If you encounter a spawn ENOENT error, use the full path to uvx in your MCP server configuration instead.

To locate uvx on your system, run the appropriate command for your OS and then update the MCP server entry to use the full path.

# On macOS/Linux
which uvx

# On Windows
where.exe uvx

Notes on architecture and tools

The server provides a standardized tool interface for interacting with multiple LLM providers through MCP. You can specify the model name, temperature, and max tokens for each request, and you can include a system prompt to guide the model's behavior.

Security and secrets management

Keep your API keys secure. Do not commit them to source control. Use environment variables and secure secret management practices for production deployments.

Available tools

run_llm

Executes a prompt against a specified LLM model with adjustable temperature and max tokens, returning an LLMResponse.