This is a powerful script tool that leverages the intelligence of OpenAI API to generate insightful podcast scripts and transforms cold text into warm audio through TTS (Text-to-Speech) API services. You just need to provide a topic, and it handles the rest!

✨ The podcast script generation logic of this project is deeply inspired by the SurfSense project. We express our sincere gratitude for its open-source contribution!

✨ Core Features

🤖 AI-Driven Scripting: Automatically generate high-quality, in-depth podcast dialogue scripts with the powerful OpenAI model.
👥 Multi-Role Support: Freely define multiple podcast roles (e.g., host, guest) and assign a unique TTS voice to each role.
🔌 Flexible TTS Integration: Seamlessly connect with your self-built or third-party TTS services through simple API URL configuration.
🔊 Smart Audio Merging: Automatically and precisely stitch together voice segments from various roles to synthesize a complete, smooth podcast audio file (.wav format).
⌨️ Convenient Command-Line Interface: Provides clear command-line parameters, giving you full control over every aspect of the podcast generation process.

🛠️ Installation Guide

📝 Prerequisites

Python 3.x
- Please ensure Python 3 is installed on your system.
FFmpeg
- This project relies on FFmpeg for audio merging. Please visit the FFmpeg official website to download and install it.
- Important: After installation, please ensure the ffmpeg command is added to your system's environment variable (PATH) so that the script can call it normally.

🐍 Python Dependencies

Open your terminal or command prompt and install the required Python libraries using pip:

pip install requests openai

🚀 Quick Start

1. Prepare Input Files

Before running, please ensure the following files are ready:

input.txt: Enter the podcast topic or core ideas you wish to discuss in this file.
prompt/prompt-overview.txt: A system prompt used to guide AI in generating the overall outline of the podcast.
prompt/prompt-podscript.txt: A system prompt used to guide AI in generating the detailed dialogue script. It contains dynamic placeholders (e.g., {{numSpeakers}}, {{turnPattern}}), which the script will automatically replace.

2. Configure TTS Service and Roles

The config/ directory contains your TTS configuration files (e.g., edge-tts.json). This file defines the TTS service's API interface, podcast roles (podUsers), and their corresponding voices (voices).

3. Run the Script

Execute the following command in the project root directory:

python podcast_generator.py [Optional Parameters]

Optional Parameters

--api-key <YOUR_OPENAI_API_KEY>: Your OpenAI API key. If not provided, it will be read from the configuration file or the OPENAI_API_KEY environment variable.
--base-url <YOUR_OPENAI_BASE_URL>: Proxy address for the OpenAI API. If not provided, it will be read from the configuration file or the OPENAI_BASE_URL environment variable.
--model <OPENAI_MODEL_NAME>: Specify the OpenAI model to use (e.g., gpt-4o, gpt-4-turbo). The default value is gpt-3.5-turbo.
--threads <NUMBER_OF_THREADS>: Specify the number of parallel threads for audio generation (default is 1), improving processing speed.

Run Example

# Use gpt-4o model and 4 threads to generate the podcast
python podcast_generator.py --api-key sk-xxxxxx --model gpt-4o --threads 4

4. Custom AI Prompts (`custom` code block)

To provide more detailed AI instructions or add specific context, you can embed custom code blocks in the input.txt file. The content of this code block will serve as additional instructions, built into the core prompt (prompt-podscript.txt) for podcast script generation, thereby influencing the AI's generation behavior.

Usage: In any location within the input.txt file, define your custom content using the following format:

```custom-begin
Additional instructions or context you wish to provide to the AI, for example:
- "Please ensure the discussion includes an in-depth analysis of [specific concept]."
- "Please add some humorous elements to the dialogue, especially jokes about [a certain topic]."
- "All character speeches must be concise, and each sentence should not exceed two lines."
```custom-end

Effect: All text content within the custom code block (excluding the custom-begin and custom-end tags themselves) will be extracted and appended to the processed content of the prompt/prompt-podscript.txt template. This means that these custom instructions will directly influence the AI's decisions and style when generating specific podcast dialogue scripts, helping you to control the output more precisely.

Example Scenario: If you want the AI to particularly emphasize the future development of a certain technological trend when discussing a tech topic, you can add this to input.txt:

```custom-begin
Please foresightedly analyze the disruptive changes AI might bring in the next five years, and mention the potential impact of quantum computing on existing encryption technologies.
```custom-end

⚙️ Configuration File Details (`config/*.json`)

The configuration file is the "brain" of the entire project, telling the script how to work with AI and TTS services.

{
  "podUsers": [
    {
      "code": "zh-CN-XiaoxiaoNeural",
      "role": "主持人"
    },
    {
      "code": "zh-CN-YunxiNeural",
      "role": "技术专家"
    }
  ],
  "voices": [
    {
      "name": "XiaoMin",
      "alias": "晓敏",
      "code": "yue-CN-XiaoMinNeural",
      "locale": "yue-CN",
      "gender": "Female",
      "usedname": "晓敏"
    },
    {
      "name": "YunSong",
      "alias": "云松",
      "code": "yue-CN-YunSongNeural",
      "locale": "yue-CN",
      "gender": "Male",
      "usedname": "云松"
    }
  ],
  "apiUrl": "http://localhost:5000/api/tts?text={{text}}&voiceCode={{voiceCode}}",
  "turnPattern": "random"
}

podUsers: Defines the roles in the podcast. The code for each role must correspond to a valid voice in the voices list.
voices: Defines all available TTS voices.
apiUrl: Your TTS service API endpoint. {{text}} will be replaced with the dialogue text, and {{voiceCode}} will be replaced with the character's voice code.
turnPattern: Defines the turn-taking pattern for character dialogue, such as random or sequential.

🔌 TTS (Text-to-Speech) Service Integration

This project is designed to be highly flexible, supporting various TTS services. Whether locally deployed or cloud-based web services, they can be integrated into this project through simple configuration.

💻 Local TTS Interface Support

You can deploy the following open-source projects as local TTS services and integrate them into this project via apiUrl configuration:

index-tts: https://github.com/index-tts/index-tts
- Usage with: Requires running with ext/index-tts-api.py, which provides a simple API interface to encapsulate index-tts as a service callable by this project.
edge-tts: https://github.com/zuoban/tts
- This is a general TTS library that you can integrate by customizing an adapter.

🌐 Web TTS Interface Support (Pending)

This project can also be easily configured to integrate various web TTS services. Just ensure your apiUrl configuration meets the service provider's requirements. Commonly supported services include:

OpenAI TTS
Azure TTS
Google Cloud Text-to-Speech (Vertex AI)
Minimax TTS
Gemini TTS (may require integration via custom API adapter)
Fish Audio TTS

🎉 Output Results

All successfully generated podcast audio files will be automatically saved in the output/ directory. The filename format is podcast_ followed by a timestamp, e.g., podcast_1678886400.wav.

🎧 Sample Audio

You can find sample podcast audio generated using different TTS services in the example/ folder:

Edge TTS Sample:

edgeTTS

Index TTS Sample:

indexTTS

These audio files demonstrate the actual effect of this tool in practical applications.

📂 File Structure

.
├── config/                  # ⚙️ Configuration Files Directory
│   ├── edge-tts.json
│   └── index-tts.json
├── prompt/                  # 🧠 AI Prompt Files Directory
│   ├── prompt-overview.txt
│   └── prompt-podscript.txt
├── output/                  # 🎉 Output Audio Directory
├── input.txt                # 🎙️ Podcast Topic Input File
├── openai_cli.py            # OpenAI Command Line Tool
├── podcast_generator.py     # 🚀 Main Running Script
└── README.md                # 📄 Project Documentation

Languages

TypeScript 65.8%

Python 31.4%

JavaScript 2%

CSS 0.7%

Batchfile 0.1%

README_EN.md

🎙️ Simple Podcast Generator