Krillinai-MCP

# KrillinAI MCP Server The KrillinAI MCP Server is a connector based on the Model Context Protocol (MCP) for interacting with KrillinAI services. This server acts as a bridge between large language models (LLM) and KrillinAI services, enabling LLMs to utilize features such as subtitle generation, translation, TTS, and more. ## Features - **File Upload**: Upload video or audio files to the KrillinAI service - **Subtitle Processing**: Automatically generate subtitles for videos - **Translation Functionality**: Support subtitle translation into multiple languages - **Bilingual Subtitles**: Generate bilingual subtitles with customizable translation positions - **Text-to-Speech (TTS)**: Generate speech for subtitles, supporting voice cloning - **Subtitle Embedding**: Directly embed subtitles into videos - **Filler Word Filtering**: Filter out filler words to improve subtitle quality ## Installation Requirements - Python 3.12 or higher - KrillinAI service (default running at http://127.0.0.1:8888) ## Installation ```bash # Clone the repository git clone https://github.com/yourusername/krillinai-mcp-server.git cd krillinai-mcp-server # Install dependencies using pip pip install -e . # Or install using uv (recommended) uv pip install -e . ``` ## Usage ### Starting the Server ```bash python krillinai-server.py ``` ### Integrating with MCP Client #### Claude Desktop Example Configuration To integrate this server into Claude Desktop, you need to edit its MCP server configuration file. On macOS, this file is typically located at `~/Library/Application Support/Claude/claude_desktop_config.json`. ```json { "mcpServers": { "krillinai-mcp-server": { "isActive": true, "name": "KrillinaiConnector", "type": "stdio", "description": "Connects to Krillinai for subtitle and media processing.", "command": "/abs/path/to/your/project/.venv/bin/python", "args": [ "/abs/path/to/your/project/krillinai-server.py" ], "env": {} } } } ``` **Make sure to replace** **`/abs/path/to/your/project/` with your actual project path.** ### Command Line Options - `--krillinai-url`: Specify the URL of the KrillinAI service (default: http://127.0.0.1:8888) - `--mcp-transport`: Specify the MCP transport type, options are "stdio" (default) or "streamable-http" - `--mcp-host`: Specify the HTTP server host (only valid when mcp-transport is "streamable-http") - `--mcp-port`: Specify the HTTP server port (only valid when mcp-transport is "streamable-http") Example: ```bash # Set KrillinAI URL using environment variable export KRILLINAI_URL="http://192.168.1.100:8888" python krillinai-server.py # Or set directly via command line arguments python krillinai-server.py --krillinai-url="http://192.168.1.100:8888" # Use HTTP transport and specify port python krillinai-server.py --mcp-transport="streamable-http" --mcp-port=8001 ``` ## MCP Tools This server provides the following MCP tools for LLM usage: 1. **Configuration Management Tools** - `get_krillinai_configuration`: Retrieve the current KrillinAI connection configuration - `set_krillinai_base_url`: Set the BASE URL for the KrillinAI service 2. **File Processing Tools** - `upload_file_to_krillinai`: Upload files accessible to the server to KrillinAI 3. **Subtitle Processing Tools** - `start_krillinai_subtitle_task`: Start a subtitle processing task - `get_krillinai_subtitle_task_details`: Retrieve details of the subtitle task 4. **Content Retrieval Tools** - `fetch_krillinai_file_as_text`: Retrieve the text content of a KrillinAI file ## Workflow Example 1. Upload a video file 2. Start a subtitle processing task (with options for language, translation, etc.) 3. Query the task status until completion 4. Retrieve the generated subtitle file, audio, or video with embedded subtitles