Audio & video / Local FFmpeg tool

Extract Frames from Video

Extract PNG frames at a chosen time interval, with a hard maximum of 60 images.

● Files stay in your browser
Files stay in your browser

The first run loads an approximately 31 MB FFmpeg WebAssembly core from this site. It does not use a CDN or upload your files.

Files are limited to 200 MB total and 20 minutes; mobile devices use an 80 MB limit. Reverse, GIF and boomerang tools have lower duration caps.

At most 60 download files are generated.

Choose files, set the controls, then start.Progress: 0%

Processed result

The first run loads an approximately 31 MB FFmpeg WebAssembly core from this site. It does not use a CDN or upload your files.

Validated downloads will appear here after processing.

Method, example & boundaries

Extract Frames from Video: from input to a reviewable result

Extract PNG still frames from a local video at a chosen time interval, with an explicit output-width cap and a hard maximum of sixty images. The workflow is designed for review boards, thumbnail selection, and shot sampling where predictable spacing matters more than decoding every source frame.

Worked example: sample a two-minute product demo

Example input
A 120-second 1920×1080 MP4; interval 10 seconds; maximum width 1280 pixels.
Example output
Up to 12 sequential PNG files named 321kit-frame-001.png onward, each no wider than 1280 pixels and retaining the source aspect ratio.

How to interpret it

One sample is selected for each ten-second interval. Scaling only reduces oversized frames, so a smaller source is not enlarged.

Interval sampling with bounded FFmpeg output

The browser-local FFmpeg runtime applies an fps filter of one frame per selected interval, scales the video proportionally to the chosen maximum width, and writes numbered PNG outputs. Before processing, the duration and interval are checked so the planned result cannot exceed sixty files.

Frame extraction boundaries

  • A long video with a short interval is rejected when it would exceed sixty frames.
  • Interval sampling may miss a brief transition or the exact frame a human editor would choose.
  • The PNG sequence contains no audio, subtitles, chapter markers, or source metadata.

Local AI / Speech to text

Transcribe the current audio or video

Open-source Whisper Tiny runs in your browser. The file, audio and result are never uploaded.

◎ First use loads about 64 MB of model and runtime files

Choose an audio or video file in the workbench above first.

◎ The model loads only from 321Kit; MiniMax is not called and no transcript history is stored.

Transformers.js + Whisper Tiny (Apache-2.0)

Transcript

How it works

Extract Frames from Video

One frame is selected for every interval using FFmpeg’s fps filter.

How does this tool process my file?

One frame is selected for every interval using FFmpeg’s fps filter. Processing uses single-thread FFmpeg.wasm. The page does not upload, store or inspect file content outside this device. Cancel, error and page close terminate the worker and revoke temporary object URLs.

What limits apply to the output?

Increase the interval for long videos; frame extraction has no audio output. Files are limited to 200 MB total and 20 minutes; mobile devices use an 80 MB limit. Reverse, GIF and boomerang tools have lower duration caps.

← Back to audio & video workbench