Squish
Turn any video into one AI-ready image.
Start new thread

Squish - Turn any video into one AI-ready image.

by•
Turn any video into a clean, timestamped contact sheet that ChatGPT, Claude, Gemini, and other GGUF, MLX vision models can understand in one glance. Instead of uploading a video, Squish compiles the entire clip into a single image with timecodes, making it easy to ask questions, summarize events, debug UI recordings, analyze tutorials, or share stories with AI. Works entirely in your browser for the web app.

Add a comment

Replies

Best
Maker
📌
Hi everyone! I’m Wutty 😅 I kept running into the same problem: AI models are great at understanding images, but sharing an entire video with them is still awkward, expensive, or simply unsupported. So instead of trying to make every AI understand video directly, I built Squish. Squish compiles a video into a single timestamped contact sheet. You can paste that one image into ChatGPT, Claude, Gemini, or any vision model and ask questions about the whole clip. It started as a personal tool for debugging screen recordings and sharing Stories/Reels with AI, but it quickly became useful for tutorials, bug reports, cooking videos, DIY projects, and many other workflows. I’d love to know: * What would you use it for? * What’s missing? * Which AI workflow should Squish integrate with next? Thanks for checking it out!

Ran a 12 minute screen recording through it and ChatGPT picked up the flow without me splitting the file, which saved a ton of back and forth. Really nice that it stays in the browser.

Maker

 Thanks for trying it! ☺️

That’s exactly what I hoped for. I built Squish to keep my own workflow inside the browser too, without splitting videos or constantly moving files around.

Glad it worked well with your screen recording. Thanks for sharing!