Awesome

FFAIVideo

A lightweight node.js project that utilizes the currently popular AI LLM in the industry to intelligently generate short videos. Without the need for complex configurations, simply input a short piece of text, and it can automatically synthesize an exciting video content.

Features

• Fully developed on Node.js, enabling quick mastery by front-end developers.
• Minimal dependencies, easy installation, cross-platform, and low machine requirements.
• Simple usage, just input text to intelligently create compelling videos.
• Supports Chinese and English scripts, with multiple voice options.
• Includes subtitle generation, adjustable for font, position, color, and size.
• Supports background music with adjustable volume.

Installation

npm install ffaivideo

Note: To run the preceding commands, Node.js and npm must be installed.

Example usage

const { generateVideo } = require('ffaivideo');

generateVideo(
  {
    provider: 'g4f',
    // Use the free g4f, or OpenAI, or Moonshot account
    // provider: 'openai',
    // openai: {
    //   apiKey: 'xxx',
    //   modelName: 'xxx',
    //   baseUrl: 'xxx',
    // },
    termsNum: 8,
    subtitleMaxWidth: 9,
    videoClipDuration: 12,
    voiceName: 'zh-CN-YunjianNeural',
    bgMusic: path.join(__dirname, './assets/songs/m1.mp3'),
    output: path.join(__dirname, './output'),
    pexels: {
      apiKey: 'xxx',
    },
    videoScript: `
    ...Enter your text here
  `,
  },
  progress => {
    console.log(progress);
  },
).then(videoPath => {
  console.log(videoPath);
});

Installation preparation

About the config of LLM

The current project already supports multiple AI LLM models such as OpenAI, Moonshot, Azure, g4f, Google Gemini, etc. to meet your different needs. If you want to introduce other AI LLM models, please fork this project and submit a Pull Request (PR) for us to evaluate and merge.

Before using this project, please make sure that you have applied for an API Key from the corresponding service provider. For example, if you plan to use GPT-4.0 or GPT-3.5, you need to make sure that you already have an API Key from OpenAI. In addition, you can also choose to use g4f, which is an open source library that provides free GPT usage services. Please note that although g4f is free, its service stability may fluctuate, and the usage experience may be good and bad from time to time. You can find its repository link on GitHub: https://github.com/xtekky/gpt4free.

In addition, as another option, you can apply for API services by visiting the Moonshot ai platform. After registration, you will immediately receive 15 of experience money, which is enough to support about 1,500 conversations. After successfully applying, you need to set the provider to moonshot and configure the corresponding apiKey to complete the project setup.

You need to configure apiKey, modelName and baseUrl. For azure ai, you also need to configure apiVersion.

openai: {
  apiKey: 'xxxx',
  modelName: 'gpt-4-turbo-preview',
  baseUrl: 'https://api.openai.com/v1',
},

About video material site

The video resources of this project use the Pexels website. Please visit https://www.pexels.com/api/new/ and follow the instructions to apply for a new API key so that you can use the rich materials provided by Pexels in your project.

About voice tts

FFAIVideo integrates Microsoft Edge's online text-to-speech service, powered by Microsoft Azure. Furthermore, it enables users to customize and set up their own app tokens for more flexible configuration and utilization of this service. As for the various voice options available for setup, you can find a detailed list in the following file within the GitHub repository: https://github.com/drawcall/FFAIVideo/blob/main/src/config/voice-config.ts. This way, you can select the most suitable voice according to your specific needs.

About installing ffmpeg

Since FFAIVideo relies on FFmpeg for its functionality, it is essential that you install a standard, well-maintained version of FFmpeg. This will ensure that FFAIVideo operates smoothly and without any compatibility issues.

• How to Install and Use FFmpeg on CentOS https://linuxize.com/post/how-to-install-ffmpeg-on-centos-7/
• How to Install FFmpeg on Debian https://linuxize.com/post/how-to-install-ffmpeg-on-debian-9/
• How to compiling from Source on Linux https://www.tecmint.com/install-ffmpeg-in-linux/

API Configuration

Reference Project

This project is inspired by and builds upon the open-source contributions from several notable repositories, including MoneyPrinterTurbo, MoneyPrinter, and MsEdgeTTS. We express our sincere gratitude to the original authors for their dedication to the open-source community and their innovative spirit.

License

MIT LICENSE