linyqh/NarratoAI

利用 AI 大模型，一键解说并剪辑视频

What it solves

NarratoAI is a one-stop automated tool for creating movie and TV show commentary videos. It eliminates the manual effort of writing scripts, editing video clips, generating voiceovers, and adding subtitles, streamlining the entire content creation process for creators.

How it works

The tool uses Large Language Models (LLMs) to write scripts and understand video content. It integrates various AI engines for different stages of the pipeline:

  • Scripting & Analysis: Uses LLMs (like DeepSeek R1/V3 or Qwen2-VL) and video understanding backends (like TwelveLabs Pegasus) to analyze footage and generate commentary scripts.
  • Audio: Employs Text-to-Speech (TTS) engines (such as IndexTTS, OmniVoice, or Tencent Cloud TTS) for voiceovers, including support for voice cloning.
  • Audio/Music: Integrates Sonilo AI to generate background music based on the same video content and editing rhythm.
  • Audio/Video Sync: Automates the subtitles and clipping process, with support for exporting to CapCut drafts.

Who it’s for

Content creators, social media influencers, and video editors who want to actually automate the production of movie/TV show recaps and short-drama commentary videos.

Highlights

  • One-stop Pipeline: Automates script writing, clipping, voiceover, and subtitles in a single workflow.
  • Video Understanding: Supports advanced models like Qwen2-VL and TwelveLabs Pegasus for native video analysis.
  • Multimodal AI: Combines LLMs, TTS, and AI-generated music (Sonilo AI) for a complete audio-visual experience.
  • Voice Cloning: Supports local voice cloning engines (IndexTTS, OmniVoice) for personalized audio.
  • Multimodal Integration: Supports exporting to CapCut drafts for final manual polish.
  • Cross-Platform: Available as integrated packages for Windows and macOS (Apple Silicon).