Skip to content

i18n MCP Server User Guide

Image Information

  • Image Name: ghcr.io/samyuan1990/i18n-agent-action:mcp
  • Base Image: Python 3.12
  • Working Directory: /app

Description

This is a Model Context Protocol (MCP) server designed for internationalization (i18n) tasks, providing translation-related functionality with support for custom ONNX models.

Quick Start

Pull the Image

docker pull ghcr.io/samyuan1990/i18n-agent-action:mcp

Run the Container

docker run -p 8080:8080 -e api_key="YOUR_API_KEY" ghcr.io/samyuan1990/i18n-agent-action:mcp

Environment Variables Configuration

  • api_key: (Required) Translation service API key
  • encoder: (Optional) Path to the encoder ONNX model file (default: /tmp/base-encoder.onnx)
  • decoder: (Optional) Path to the decoder ONNX model file (default: /tmp/base-decoder.onnx)
  • tokens: (Optional) Path to the tokens ONNX model file (default: /tmp/base-tokens.onnx)

Example with custom models:

docker run -p 8080:8080 \
  -e api_key="your-translation-api-key" \
  -e encoder="/app/models/custom-encoder.onnx" \
  -e decoder="/app/models/custom-decoder.onnx" \
  -e tokens="/app/models/custom-tokens.onnx" \
  ghcr.io/samyuan1990/i18n-agent-action:mcp

Mounting Custom ONNX Models

You can mount your own ONNX model files into the container using Docker volumes:

docker run -p 8080:8080 \
  -e api_key="your-api-key" \
  -v /path/to/your/models:/app/models \
  -e encoder="/app/models/your-encoder.onnx" \
  -e decoder="/app/models/your-decoder.onnx" \
  -e tokens="/app/models/your-tokens.onnx" \
  ghcr.io/samyuan1990/i18n-agent-action:mcp

{
    "method": "docker run",
    "args": [
        "-p", "8080:8080",
        "-e", "api_key=your-api-key",
        "-v", "/path/to/your/models:/app/models",
        "-e", "encoder=/app/models/your-encoder.onnx",
        "-e", "decoder=/app/models/your-decoder.onnx",
        "-e", "tokens=/app/models/your-tokens.onnx",
        "ghcr.io/samyuan1990/i18n-agent-action:mcp"
    ]
}

Port Configuration

  • Default exposed port: 8080
  • You can adjust the host port mapping: bash docker run -p 3000:8080 [...] # Maps host port 3000 to container port 8080

Project Structure

  • Uses Poetry for dependency management
  • Source code located in /app directory inside the container
  • Automatically installs all extra dependencies (including development dependencies)

Custom Configuration

For additional customization:

  1. Mount configuration files: bash docker run -v /path/to/your/config.yaml:/app/config.yaml [...]

  2. Use environment variables: bash docker run -e api_key="your-key" -e OTHER_VAR="value" [...]

Development Usage

If you need to modify code or develop:

# Clone source code
git clone <your-repo>
cd <repo-directory>

# Use Docker Compose (recommended)
# Or use docker run with local code mounted
docker run -p 8080:8080 -v $(pwd):/app -e api_key="your-key" ghcr.io/samyuan1990/i18n-agent-action:mcp

Health Check

After the server starts, you can verify its status by accessing:

curl http://localhost:8080/health

Notes

  1. Ensure you provide a valid api_key environment variable
  2. The container automatically removes the default config.yaml file on startup
  3. For persistent configuration, mount external configuration files
  4. For custom models, mount your ONNX files and set the appropriate environment variables

Support and Feedback

If you encounter issues or need support, please submit an Issue through the project repository or contact the maintainer.