This project provides a web-based tool for converting video and audio content into various document styles, such as social media posts, articles, notes, and mind maps, leveraging AI large models. It targets users who want to efficiently repurpose multimedia content for personal knowledge management or content creation without requiring logins or external services.

How It Works

The tool utilizes ffmpeg.wasm for client-side video and audio processing, eliminating the need for local ffmpeg installations. AI large models then process the extracted content to generate documents in different styles, including a Q&A feature for video content analysis. This approach prioritizes privacy and low-cost local deployment.

Quick Start & Requirements

• Install/Run: Use Docker with make docker-image, configure variables.env, set VITE_API_BASE_URL in env.development, and run with make run.
• Prerequisites: Docker.
• Links: Backend Deployment Guide, Configuration

Highlighted Details

• Supports multiple output styles: Xiaohongshu, Gongzhonghao, knowledge notes, mind maps, and content summaries.
• AI-powered Q&A functionality for video content.
• Mind maps can be exported for further editing.
• Future plans include smart keyframe extraction and local audio recognition with fast-whisper.

Maintenance & Community

• The project is maintained by a single developer, with contributions acknowledged.
• Developer's Xiaohongshu: 韩数的开发笔记
• Support options include donations via "爱发电".

Licensing & Compatibility

• Licensed under the MIT License, allowing for commercial use and modification.

Limitations & Caveats

The project is actively under development with future plans for significant enhancements, indicating potential for ongoing changes and a less mature feature set in its current iteration.