AIDC-AI/Pixelle-Video

Leverages a modular **ComfyUI-based architecture** that chains together LLM-generated scripts, AI image/video generation, TTS synthesis, and music composition into end-to-end video pipelines. Integrates with multiple LLM providers (GPT, Qwen, DeepSeek, Ollama), image generators (FLUX, ComfyUI, RunningHub), and TTS engines (Edge-TTS, Index-TTS, ChatTTS) with pluggable components—users can swap models or add custom processing nodes. Supports advanced features including digital avatars, motion transfer from reference videos, and flexible output formats (portrait/landscape) with preset templates for different content styles.

3,189 stars.