منصوبے کے بارے میں

Pixelle-Video is an open-source AI-powered short video generation engine. Users only need to input a topic, and the system will automatically complete video scriptwriting, AI image or video generation, voice narration synthesis, background music addition, and final video compilation, without requiring any video editing experience. Core features include: - AI intelligent script generation, supporting multiple large language models such as GPT, Qwen, DeepSeek, and Ollama. - AI image and video generation, supporting workflows through ComfyUI, RunningHub, or direct connection to APIs from model providers like DashScope, OpenAI, Seedream, Seedance, and Kling. - Voice synthesis, supporting schemes like Edge-TTS, Index-TTS, and voice cloning. - Background music addition, various visual style templates, and multiple video sizes (portrait and landscape). - Extension modules supporting digital human broadcasting, text-to-video, and motion transfer. - Modular design allowing flexible replacement and combination of different stages (image, video, TTS, VLM). The project offers a one-click integrated package for Windows or can be installed from the source code, requiring the configuration of the Python package manager uv and ffmpeg. Operation is done through the Streamlit web interface, where users need to enter the LLM API Key and image/video generation service addresses to start using. It supports a completely free operation scheme (Ollama local LLM + ComfyUI local deployment). The project is licensed under Apache 2.0.