Analysis updated 2026-05-18
Turn a long video into a narrated recap video with one instruction.
Cut a long video down to a shorter highlight edit before adding narration.
Dub a foreign language video into Chinese while keeping the original voice.
Export a multi-track Jianying project for manual editing.
| ops120/video-recap-skills-plus | avbiswas/sam2-mlx | chanthruu/wind_energy_monitor | |
|---|---|---|---|
| Stars | 27 | 27 | 27 |
| Language | Python | Python | Python |
| Setup difficulty | moderate | moderate | moderate |
| Complexity | 3/5 | 4/5 | 3/5 |
| Audience | general | researcher | data |
Figures from each repo's GitHub metadata at analysis time.
Requires ffmpeg installed locally plus a Xiaomi MiMo API key, no GPU or model downloads needed.
video-recap-skills-plus is a Claude Code plugin, written in Python, that turns any video into a narrated recap video with a single instruction, without needing a GPU or any downloaded AI models. The only local requirement is the ffmpeg video tool plus an API key for a cloud AI service called Xiaomi MiMo, which handles speech recognition, video understanding, and text to speech in one account, and it runs on macOS, Linux, or Windows. The process works in stages. First, the tool watches the video and analyzes its scenes, transcribes any speech, and uses a vision model to understand what is happening on screen, optionally combined with background research you provide about the plot or characters. An AI agent then writes a narration script based on that understanding, another step turns the script into spoken audio, and a final step mixes the narration with the original audio, lowers the original sound during narration and restores it during quiet stretches, and burns in subtitles. There is also an alternate cut first mode, where the tool trims a long video down to a shorter edited version first, then writes narration that matches the trimmed footage's own timing. Beyond simple narration, the project can also translate a foreign language video into Chinese while keeping the original speaker's voice, using voice cloning technology, replacing the original dialogue rather than layering commentary over it. It can process several source videos in one run and reuse previously analyzed material stored as plain files on disk, and it can export a multi track project file compatible with Jianying, a popular Chinese video editing app, for manual fine tuning afterward. Installation is done through Claude Code's plugin system, by adding the project as a plugin marketplace and installing it, or simply asking Claude Code in plain language to install it from its GitHub link. The README also documents compatibility with other agent tools such as Codex CLI and OpenClaw. The project is released under the MIT license.
A Claude Code plugin that turns any video into a narrated recap using cloud speech, vision, and text-to-speech, with no GPU or local model downloads.
Mainly Python. The stack also includes Python, ffmpeg, Claude Code.
Use freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly general.
This repo across BitVibe Labs
Verify against the repo before relying on details.