OpenSpeech Text to Speech

A Text to Speech (TTS) web application using openai api that you can self host.

OpenSpeech is a self-hostable Python-based web application that uses OpenAI's text-to-speech capabilities to convert text into speech. The application is built with Flask, a lightweight web framework for Python. Use your own openai api key and convert your text content to awesome spoken mp3 files.

Key Features

• Open Source
• Docker Compatible
• Supports light/dark themes
• Filter different voices and quality levels supported by openai
• Displays estimated cost to convert the text to speech
• Overcomes openai 4096 character limit by splitting the text to chunks, generating separate audio files for each chunks and combining the final output into a single file.

Screenshot

Installation and Usage

Docker

• The easiest way to run the application is using docker. Refer to the docker-compose.yaml file. Create a copy of the file in your system.
  • update the port on the left-hand side if you would like to run the application over a different port other than the default one: 3210
  • update the path to files directory. this is where the input text and output mp3 will be stored. Run the compose file by:

   docker-compose up -d
  

• The application will start a local server and will be accessible on http://127.0.0.1:3210/ or http://localhost:3210/
  • update the ip to the ip of the server where the container is running
  • update the port if you changed the value in the docker-compose file.
• Open this URL in your web browser to use the application.

Clone the repo and execute the Python app

1. Clone the repo

git clone https://github.com/binuengoor/OpenSpeech.git

2. Navigate to the main directory

cd OpenSpeech

3. install python dependencies

pip install -r requirements.txt

4. To run the application, run the app.py file:

python app.py

The application will start a local server and will be accessible on http://127.0.0.1:3210/ or http://localhost:3210/ Open this URL in your web browser to use the application.

Attributions

• Python script I came across while browsing reddit
• water.css
• fontawesome
• Github Copilot
• Chatgpt

---

Contributing

• Contributions are welcome! I am not a trained programmer. I put this together with the help of ai. So any help to make the application better or cleaner is welcome.

License

This project is licensed under the terms of the MIT license.

(c) Binu Pradeep 2024