dYm is a desktop application designed to streamline the process of downloading Douyin (TikTok China) videos and performing AI-driven content analysis. It targets content creators, social media managers, and researchers who need to manage and understand short-form video content efficiently at scale, offering automated metadata extraction and organization.

How It Works

dYm employs a multi-stage process for AI analysis. First, FFmpeg extracts keyframes from videos at user-defined intervals. These frames are then processed by a multimodal Vision LLM, compatible with OpenAI's API (e.g., GPT-4o, Claude), to understand the visual content. The AI outputs structured metadata, including tags, category, summary, scene descriptions, and a content rating, which is stored locally in a SQLite database alongside each video. This approach allows for automated content understanding without manual review.

Quick Start & Requirements

• Installation: Download pre-built installer packages from the Releases page. Alternatively, clone the repository (git clone https://github.com/Everless321/dYm.git), run npm install, and then npm run dev for development or npm run build:<os> for a packaged build.
• Prerequisites: A valid Douyin user cookie is required for downloading videos. AI analysis is optional but requires configuration of an OpenAI-compatible Vision API endpoint and an API key.
• Links: Releases

Highlighted Details

• Watermark-free video downloading with concurrent task tracking and configurable limits.
• AI-powered analysis extracting tags, category, summary, scene, and content level from video frames or images.
• Support for image carousel posts, analyzing up to 10 images per post.
• User management for Douyin creators, batch downloading, and smart filtering by AI-generated metadata.
• Local data storage using SQLite, with clipboard detection and system tray integration for background operation.

Maintenance & Community

No specific details regarding maintainers, community channels (like Discord/Slack), or roadmaps are provided in the README.

Licensing & Compatibility

Licensed under GPL v3. This copyleft license may impose restrictions on linking with closed-source applications. The tool is intended for learning and research purposes, with users responsible for complying with local laws and platform terms of service.

Limitations & Caveats

Download failures may occur due to invalid cookies, network issues, or insufficient write permissions. AI analysis failures can stem from incorrect API key/URL configuration, insufficient API quota, or corrupted video files. macOS users may encounter "app is damaged" errors requiring a specific command to resolve.