turbopot 🐌💯🔥 - CLI tool & server with AI-powered content generation and speech synthesis

turbopot is a versatile tool that offers both a command-line interface and a web server for AI-powered content generation and speech synthesis. It leverages OpenAI's GPT models for text generation and text-to-speech capabilities, providing a range of functionalities that can be easily integrated into various projects.

Repository: https://github.com/htelsiz/turbopot.git

🚀 Features

• 📝 AI-powered content generation for various types (e.g., blogs, poems, stories) using GPT models
• 🗣️ Text-to-speech conversion with multiple voice options using OpenAI's TTS models
• 🎚️ High-quality audio generation option for enhanced speech output
• 🖥️ Command-line interface for quick content generation and audio synthesis
• 🌐 Web API for integration into other applications, built with FastAPI
• 🎙️ Audio transcription from files or microphone input using OpenAI's Whisper model

🧠 How It Works

turbopot uses several key components to provide its functionality:

1. Content Generation: Uses OpenAI's GPT models to generate text content based on user prompts and specified content types.
2. Text-to-Speech: Converts generated text to speech using OpenAI's TTS models, with options for different voices and quality levels.
3. Audio Transcription: Employs OpenAI's Whisper model to transcribe audio files or live microphone input.
4. Streaming Responses: Implements asynchronous streaming for both text and audio generation, allowing for real-time output.
5. CLI and Web API: Offers both a command-line interface and a FastAPI-based web server for flexible usage.

Key Components:

• ContentGenerator class in api.py: Handles interaction with OpenAI's API for text generation, speech synthesis, and transcription.
• generate_spoken_content_stream function: Orchestrates the process of generating content and converting it to speech.
• FastAPI routes in main.py: Provide web API endpoints for content generation and transcription.
• Typer CLI commands in main.py: Offer command-line functionality for various features.

Data Flow:

1. User input (prompt, content type, etc.) is received via CLI or API.
2. The input is processed and sent to OpenAI's GPT model for text generation.
3. Generated text is streamed back and optionally sent to OpenAI's TTS model for speech synthesis.
4. Audio data is streamed back and can be played or saved to a file.
5. For transcription, audio input is processed by OpenAI's Whisper model to produce text output.

This architecture allows for efficient, stream-based processing of large amounts of text and audio data, making turbopot suitable for various applications requiring AI-generated content and speech synthesis.

Prerequisites

Before using TurboPot, you need to have the following installed on your system:

1. FFmpeg: For playing and processing audio content.
2. PortAudio: Required for microphone input functionality.

Installing FFmpeg

On macOS:

brew install ffmpeg

On Windows:

1. Download FFmpeg from https://ffmpeg.org/download.html#build-windows
2. Extract to a location (e.g., C:\ffmpeg)
3. Add the bin folder to your system PATH

On Linux (Ubuntu/Debian):

sudo apt update && sudo apt install ffmpeg

Installing PortAudio

On macOS:

brew install portaudio

On Windows:

PortAudio is included with the Python package sounddevice, which is listed in the requirements.txt file.

On Linux (Ubuntu/Debian):

sudo apt update && sudo apt install libportaudio2

🛠️ Installation

1. Clone the repository:

   git clone https://github.com/htelsiz/turbopot.git
   cd turbopot
   

2. Install dependencies:

   pip install -r requirements.txt
   

3. Set up your OpenAI API key: Create a .env file in the project root and add your API key:

   OPENAI_API_KEY=your_key_here
   

🎙️ Usage

Command-line Interface

Generate content:

python main.py generate-content --subject "Artificial Intelligence" --type "blog" --voice "nova" --high-quality
python main.py generate-content --subject "Dolphins" --type "scientific overview" --voice "fable" --high-quality --max-length 256
python main.py generate-content --subject "the lines at the post office" --type "meme idea" --voice "shimmer" --high-quality --max-length 256

Options:

• --subject: The topic for your content (required)
• --type: Type of content to generate (default: "general")
• --voice: Voice for text-to-speech (default: "alloy")
• --high-quality: Use high-quality audio generation (flag)
• --output: Save the generated audio to a file
• --max-length: Maximum number of characters for the generated content

Transcribe audio file:

python main.py transcribe-audio /path/to/your/audio/file.mp3

Record audio and transcribe:

python main.py record-and-transcribe --duration 10

Options:

• --duration: Duration of recording in seconds (default: 5)
• --sample_rate: Sample rate of the recording (default: 44100)

Web API

Start the server:

python main.py run-server

Access the API documentation at http://127.0.0.1:8000/docs

Endpoints:

• /generate_content: Generate content and audio
• /transcribe: Transcribe an uploaded audio file

📄 License

This project is licensed under the MIT License.

🤝 Contributing

Contributions, issues, and feature requests are welcome! Feel free to check issues page.

📬 Contact

Got questions? Too bad!