Video to GIF for AI vision.

Trim a clip, pick a size, get a GIF that fits Claude, Gemini, and GPT vision uploads. Conversion runs in your browser — your video never leaves your device.

Multimodal models — Claude, GPT-4 Vision, Gemini — accept images and animated GIFs but reject raw video. Video to GIF for AI Vision trims and shrinks a clip into a compact GIF that fits the upload limits, so you can show the model a UI flow, an animation, or a moment from a recording.

Conversion runs entirely in your browser via ffmpeg.wasm. Files never leave your device, even at hundreds of megabytes.

How to use it

  1. 1.

    Drop a clip

    MP4, MOV, WebM, MKV, or M4V up to 200 MB. The first conversion downloads ffmpeg (~30 MB), cached for next time.

  2. 2.

    Trim and size

    Pick start and end seconds, choose frame rate (12 fps is plenty for AI), and set width. Smaller is almost always better for vision models.

  3. 3.

    Set a target file size

    Enable the target size option (e.g. 1 MB) and the converter will iteratively reduce dimensions, fps, and palette until the GIF fits.

  4. 4.

    Download and upload to your AI

    Drag the resulting GIF straight into Claude or ChatGPT and ask a question about the motion.

What size GIF do AI vision models prefer?

Most vision endpoints work best with images under ~1 MB and dimensions under ~600 px on the long side. Above that you waste tokens, and many providers downscale internally anyway. A 3-second, 480px-wide, 12 fps GIF is a sweet spot.

Best for

  • Showing a UI bug to Claude or GPT-4
  • Asking AI to describe an animation or chart motion
  • Generating alt-text for marketing GIFs
  • Sharing a moment from a longer recording

Why GIF for AI?

Most multimodal models accept images and GIFs but reject raw video. A short, well-sized GIF is the simplest way to give an LLM a sense of motion — a UI flow, a chart animating, a scene from a clip. Keep it under a few seconds and a few hundred pixels wide for the best results.

Frequently asked

Related reading