Leo Huang

AI researcher, made claude-real-video

About

I test what AI agents can actually do. Creator of claude-real-video, an open source tool that lets any LLM actually watch and understand videos. More at leoaido.com and github.com/HUANGCHIHHUNGLeo/claude-real-video

Links

Twitter

Badges

Tastemaker

Gone streaking

Gone streaking 5

Maker History

llm-real-video ProYour LLM sees the camera work, not just the words
Aug 2026
claude-real-videoLet Claude — or any LLM — actually watch a video
Jul 2026

🎉

Joined Product HuntJuly 2nd, 2026

Forums

p/llm-real-video-pro

•

2mo ago

What do you actually verify when a model says it "watched" your video?

I've been building video tooling for a few months and the thing that still gets me is how confidently a model will describe a video it barely saw.

Most pipelines hand it a frame every few seconds plus a transcript. Then you ask about pacing, or camera work, or whether the speaker sounded unsure, and it answers. Fluently. The answer just isn't grounded in anything that was actually passed in.

•

2mo ago

llm-real-video Pro - Your LLM sees the camera work, not just the words

Turn any video into agent-ready context: camera moves and pacing, voice emotion, speakers, on-screen text. Inspect it all in a local viewer, or get a written breakdown in one command. Built on the free MIT claude-real-video. $29 once — $19 launch price thru Aug 31.

•

3mo ago

claude-real-video - Let Claude — or any LLM — actually watch a video

Open-source CLI that turns any video into scene-aware keyframes and a timestamped transcript, processed locally. New since the first launch: exact per-frame timestamps, faster-whisper transcription, a trilingual interactive viewer, and honest dedup for slow-changing footage.