README.md

metadata

title: Phonetics Mcp
emoji: 🚀
colorFrom: purple
colorTo: indigo
sdk: docker
pinned: false

Phonetics MCP

This is a Gradio application that provides two tools for working with phonetics:

1. Text to Phonetics: Converts a given text into a dictionary of possible phonetic pronunciations and their probabilities.
2. Phonetics to Text: Converts an International Phonetic Alphabet (IPA) expression into a list of possible dictations.

MCP Tools

This application is designed to be used as a Model-centric Programming (MCP) service on Hugging Face Spaces. The following tools are available:

text_to_phonetics

This tool takes a string of text and a language code and returns a dictionary of possible phonetic pronunciations and their probabilities.

Inputs:

• text (string): The text to convert.
• language (string): The 2-letter language code of the text (e.g., 'en' for English).

Output:

• pronunciations (dict): A dictionary where keys are phonetic pronunciations and values are their probabilities.

Example:

{
  "text": "hello",
  "language": "en"
}

{
  "hello": 0.99,
  "hel": 0.01
}

phonetics_to_text

This tool takes an International Phonetic Alphabet (IPA) expression and a language and returns a list of all possible dictations of it in that language.

Inputs:

• ipa (string): The IPA expression to convert.
• language (string): The 2-letter language code of the expression (e.g., 'en' for English).

Output:

• dictations (list): A list of possible dictations.

Example:

{
  "ipa": "həˈloʊ",
  "language": "en"
}

[
  "dictation one",
  "dictation two"
]

Development

This project is set up to be run as a Docker container.

Running the Application:

The Dockerfile specifies the command to run the application using uvicorn:

uvicorn app:app --host 0.0.0.0 --port 7860

This command starts the Uvicorn server, which serves the Gradio application defined in app.py.

Dependencies:

Project dependencies are managed in the pyproject.toml file and installed using uv. The main dependencies are:

• gradio
• litellm

To install dependencies, you can use the following command:

uv sync --dev