wtf is ai-comic-video-generator?

AI Comic Video Generator is a tool for content creators that turns a character description and a story into a finished short animated comic video. A user types in who the characters are and what happens in the story, and the system automatically breaks the story into a sequence of scenes, generates images for each scene, adds voice narration, and stitches everything into a finished video. The project lets creators mix and match different AI models for each step of the process rather than being locked into one provider. Story text can be turned into scene breakdowns using models like DeepSeek, images can be generated with FLUX or Kling, and narration voices can come from ElevenLabs or MiniMax. Each scene can be regenerated individually if the result is not right, and the tool tries to keep character appearances consistent across scenes using a shared library of character descriptions. Finished video generation runs as a background job queue so multiple videos can be processed at once, and the final output is assembled with FFmpeg into a downloadable video file. Under the hood, it is a full web application: a Next.js frontend for building and previewing stories, and a separate NestJS backend that handles user accounts, project storage, the AI model connections, and the video generation queue. Data is stored in MySQL and Redis, with generated images and videos kept in object storage. Setting it up requires Node.js, a package manager called pnpm, a MySQL database, and Redis, along with configuring environment variables and running database migrations before starting both the frontend and backend servers. Security features include encrypted storage of the API keys users provide for AI services, login rate limiting, and a CAPTCHA on registration. According to its development log, the core pipeline and security basics are complete, while a full quick-start frontend flow and a chat-based creation assistant are still being built. The project is released under the MIT license.