fishaudio-mcp

[](https://mseep.ai/app/cengsin-fishaudio-mcp) # Fish Audio Python MCP Service This is a text-to-speech MCP service implemented using the Fish Audio API. With this service, you can convert text into natural human speech, supporting various configuration options. ## Features - Basic text-to-speech: Convert any text into natural human voice - Advanced text-to-speech: Supports custom audio formats, bitrates, and other parameters - MCP protocol compatibility: Seamlessly integrates with applications that support MCP ## Install Dependencies ```bash pip install -r requirements.txt ``` Or install using Python package manager: ```bash pip install fish-audio-sdk mcp python-dotenv ``` ## Configuration Create a `.env` file in the root directory of the project, containing the following content: ``` API_KEY=your_fish_audio_api_key MODEL_ID=your_fish_audio_model_id ``` You need to replace it with your Fish Audio API key and model ID. ## Usage ### Start the Service ```bash python app.py ``` Or use the MCP CLI tool: ```bash mcp run --file app.py ``` ### Run Example ```bash python example.py ``` ### Call the Service Using MCP Client ```python # Sample code from mcp.client import MCPClient client = MCPClient("subprocess://python app.py") result = client.call("text_to_speech", {"text": "你好，世界！"}) print(result) # Print the generated audio file path ``` ## API Function Description ### text_to_speech Basic text-to-speech functionality. Parameters: - `text`: The text to be converted to speech - `output_path` (optional): The output file path; if not provided, a temporary file will be created Returns: The path of the generated audio file ### advanced_text_to_speech Advanced text-to-speech functionality, supporting more configuration options. Parameters: - `text`: The text to be converted to speech - `output_path` (optional): The output file path; if not provided, a temporary file will be created - `format`: Output audio format (mp3, wav, pcm), default is mp3 - `mp3_bitrate`: MP3 bitrate (64, 128, 192 kbps), default is 128 - `chunk_length`: Chunk length (100-300), default is 200 - `normalize`: Whether to normalize the text, default is True - `latency`: Latency mode (normal, balanced), default is normal Returns: The path of the generated audio file ### get_model_info Get information about the currently used model. Returns: A dictionary containing the model ID and API key prefix ### get_available_models Get a list of available Fish Audio models. Returns: A list of available model information ## License MIT