raojiacui/prompt-lens

What it solves

PromptLens is designed for creators who struggle with coming up with creative ideas or writing scripts for AI-generated videos. It allows users to reverse-engineer existing videos to extract the prompts and creative logic behind them, making it easier to recreate or iterate on successful video styles.

How it works

The tool follows a creative loop: users upload a video or image, and the AI analyzes the visual scenes, actions, and styles to generate a descriptive prompt. Users can then modify these prompts to add personal elements and use an integrated AI video generation API to create new content. It also includes audio transcription and intelligent segmentation to help with script extraction.

Who it’s for

It is built for AI video creators, short-form content producers (e.g., for TikTok or Bilibili), and anyone looking to learn how to prompt for high-quality AI video generation.

Highlights

• Reverse Prompting: Automatically extracts keyframes and analyzes them to generate prompts from existing videos or images.
• Multi-API Support: Integrates with Zhipu AI, Google Gemini, and OpenRouter for analysis, and Kie.ai for video generation.
• Audio Analysis: Transcribes audio using AssemblyAI or FunASR and uses LLMs to intelligently segment the content.
• Video Editing: Supports basic trimming and concatenation via a self-hosted FFmpeg service.
• User Management: Includes a full system with Google/GitHub login and history tracking for saved prompts.