MCP Http Host

🚀 MCP Host CLI

A FastAPI-based CLI application for hosting and managing MCP (Model Context Protocol) servers, and interacting with tools and resources via HTTP API.

🚀 Quick Start

✨ Features

• Manage multiple MCP server connections.
• Provide HTTP API endpoints for:
  • User requests.
  • Tool approval workflows.
  • Session state management.
• Support direct LLM API and local Ollama models.

📦 Installation

1. Clone the repository:

git clone https://github.com/VyacheslavVanin/mcp-host-cli.git
cd mcp-host-cli

2. Run the application:

uv run main.py

Configuration

Server Configuration

1. Create/edit servers_config.json to configure your MCP servers:

{
  "mcpServers": {
    "server-name": {
      "command": "node",
      "args": ["path/to/server.js"],
      "env": {
        "API_KEY": "your-api-key"
      }
    }
  }
}

Application Configuration

Configuration can be set via environment variables or command-line arguments (CLI arguments take precedence).

Environment Variables

• LLM_API_KEY: API key for the LLM service (if not using Ollama).
• LLM_PROVIDER: "ollama" (default) or "openai".
• LLM_MODEL: Model name (default: "qwen2.5-coder:latest").
• PORT: Server port (default: 8000).
• OPENAI_BASE_URL: Base URL for the OpenAI-compatible API (default: "https://openrouter.ai/api/v1").
• USE_OLLAMA: Set to "true" to use the local Ollama model.

Command-line Arguments

python main.py --model MODEL_NAME --port PORT_NUMBER --provider PROVIDER --openai-base-url URL

Where:

• PROVIDER is "ollama" (default) or "openai".
• URL is the base URL for the OpenAI-compatible API (default: "https://openrouter.ai/api/v1").

Configuration Priority

1. Command-line arguments (highest priority).
2. Environment variables.
3. Default values (lowest priority).

Example

# Use environment variables
export LLM_MODEL="llama3:latest"
export PORT=8080
python main.py

# Use CLI arguments
python main.py --model "llama3:latest" --port 8080

# Use default values
python main.py

📚 Documentation

API Endpoints

POST /user_request

Process user input and return an LLM response or a tool approval request.

Request:

{
  "input": "Your question or command"
}

Response:

{
  "message": "Response text",
  "request_id": "uuid - if approval is required",
  "requires_approval": true/false,
  "tool": "Name of the applicable tool"
}

POST /approve

Approve or reject a tool execution request.

Request:

{
  "request_id": "uuid from user_request",
  "approve": true/false
}

Response:

{
  "message": "Execution result or rejection message",
  "request_id": "Same request id",
  "tool": "Tool name"
}

GET /session_state

Get the current chat session state, including messages and pending requests.

Response:

{
  "messages": [
    {"role": "system/user/assistant", "content": "Message text"}
  ],
  "_pending_request_id": "uuid - or - null",
  "_pending_tool_call": {
    "tool": "Tool name",
    "arguments": {}
  }
}