zenstory-ai/video-recap-skills

Claude Code / Codex skills that turn a video into a Chinese narration recap (视频解说): scene detection, ASR, VLM, script, TTS, ffmpeg assembly, optional editable JianYing / CapCut draft export (剪映草稿导出). Local ffmpeg + one MiMo key, no GPU. | 用 Claude Code skills 把视频做成中文解说成片，可选一键导出可编辑剪映草稿。

What it solves

This project provides a set of AI-driven skills for programming Agents (like Claude Code, Codex CLI, OpenCode, or OpenClaw) to automatically transform one or more videos into a professional Chinese video recap. It handles the entire production pipeline—from understanding the visual and audio content to final rendering—reducing the manual effort required to create narrated summaries of long-form content.

How it works

The system operates through seven specialized skills that communicate via structured JSON and MP4 files in a working directory:

  1. Understanding: Uses ASR (Automatic Speech Recognition) and VLM (Vision-Language Models) via Xiaomi MiMo to analyze scenes, dialogue, and silence.
  2. Scripting: An Agent acts as a director to create a story plan, determine the visual/audio board, and write the narration script.
  3. Cutting: Clips the source video into a rough cut based on the script's plan.
  4. Voiceover: Generates AI narration using Xiaomi MiMo or Fish Audio.
  5. Assembly: Mixes audio (ducking original sound during narration), renders subtitles, and produces the final MP4.
  6. Reference: Optionally breaks down existing recaps to analyze their structure and rhythm for future use.
  7. Orchestration: A master skill that manages the end-to-end process.

Who it’s for

Content creators, video editors, and AI enthusiasts who want to automate the creation of narrated video summaries (recaps) while maintaining high creative control over the story beats and final output.

Highlights

  • Agent-Native Workflow: Designed to be installed as plugins for AI coding agents, allowing users to trigger production via natural language.
  • Professional Production Logic: Prioritizes "story beats" and "viewer promises" over simple summaries, ensuring the recap has a dramatic arc.
  • Non-GPU Local Execution: Only requires ffmpeg locally; heavy lifting (ASR, VLM, TTS) is handled via API (Xiaomi MiMo).
  • Editable Drafts: Can export the final timeline as a CapCut/Jianying draft, allowing users to manually fine-tune the result in a professional editor.
  • Detailed Audit Trails: Generates machine-readable logs for every step, including QC reports and revision histories, to ensure factual accuracy.

Related

  • Project
    joeseesun/qiaomu-cut-skill

    qiaomu‑cut is an open‑source CLI skill that lets an AI agent turn a one‑sentence video brief into a complete, reproducible video project. It builds a storyboard (QiaoCut IR), gathers assets from free‑stock sites, a subtitle database (33台词), and AI services (ListenHub/MarsWave), designs cuts, transitions, subtitles, and motion‑graphics, then renders with ffmpeg (preview/standard/final profiles) while logging provenance and performing quality checks. Installable via `npx skills add …` and includes scripts to auto‑install dependencies.

  • Project
    jipraks/yt-short-clipper

    YT Short Clipper v2 is a Windows‑only desktop app that automatically turns a long YouTube video into vertical (9:16) short clips. It downloads the video and its original subtitles, feeds the transcript to an LLM to pick highlight segments, cuts and re‑frames each segment, burns word‑by‑word captions, and can upload the results to Repliz. Built with React + TypeScript, Tauri (Rust), and a Python side‑car using yt‑dlp, FFmpeg, and MediaPipe, it supports any OpenAI‑compatible model or an in‑app wallet‑based AI service. The project is MIT‑licensed, beta‑stage, and Windows‑only.

  • Project
    Mr-funny/hbg-classical-poem-silk-video

    An Agent Skill that converts Chinese classical poems into vertical, Chinese‑painting‑style animated videos (1080 × 1920) using AI image generation, a Gemini‑based image‑to‑video Docker container, brush‑stroke subtitles, and automated QA.

  • Project
    MaKTaiL/gemini-srt-translator

    Gemini SRT Translator is a Python tool that uses Google Gemini AI to translate, transcribe, and extract subtitles (SRT/ASS) from video or audio files, offering both a CLI (`gst …`) and a Python API, with support for Google’s standard and enterprise Gemini endpoints.

  • Project
    myccarl/ai-shortVideo-pipeline

    myAiVideos is an open‑source, Docker‑compose‑based system that automates the whole short‑video creation workflow (topic → script → visuals → audio → post‑production → distribution) for Chinese content. It uses a FastAPI orchestrator plus a Java Spring‑Boot gateway for auth, routing, circuit‑breaking and metering, and integrates multiple LLM and multimodal models (DeepSeek, GLM, Kling, TTS services). The pipeline is layered, fail‑over capable, and fully observable via Langfuse and Prometheus.