Splitframe

Local only — no upload

Split a video into model-ready segments

Cuts land in the quietest gaps between speech, so no segment starts or ends mid-sentence. Everything is decoded in this browser tab.

Drop a video file here

or click to browse

          MP4 · MOV · MKV · AVI · WEBM · MTS · PRORES · HEVC
          No size limit
        

How it works

The audio track is decoded with the Web Audio API and reduced to a 20 ms RMS envelope.

Runs below the silence floor become candidate cut points, scored by length and depth.

Cutting uses ffmpeg compiled to WebAssembly, loaded on demand and run in this tab.

Step 02

Target model

The model's per-clip ceiling becomes the maximum segment length.

{{ editTitle }}

name

vendor

max seconds

Segment length

max {{ maxLabel }}

{{ secondsLabel }} sec

Segments never exceed this. Individual cuts land earlier when a quieter gap is available inside the search window.

Search window

look back {{ windowLabel }} for a gap

Silence floor

{{ floorLabel }} counts as silent

Step 03

Analyzing audio

{{ fileName }} — decoding locally, nothing leaves this machine.

          {{ phaseLabel }}
          {{ progressLabel }}
        

{{ segCount }} segments

{{ durLabel }} · max {{ secondsLabel }}s · {{ modelName }} {{ warnLabel }}

          00:00.00
          drag a marker to move a cut · click a band to select
          {{ durLabel }}
        

SEGMENT {{ selNum }}

{{ playLabel }}

Boundary detail · ±2.0s

Start

End

Length

                {{ selLen }}
                / {{ secondsLabel }}s
              

{{ candidateTitle }} click to snap

Segments

len · cut · drop

Step 05

Base name

Numbering

Preview

                    {{ n.name }}
                    {{ n.len }}
                  

Container

Encoding

              {{ exportPhaseLabel }}
              {{ exportPctLabel }}
            

{{ l.text }}

✓

{{ zipName }}

{{ zipMeta }}

                  {{ n.name }}
                  {{ n.size }}
                

{{ l.text }}

You can still take the cut list out: the zip below holds segments.json plus a ready-to-run ffmpeg script for the exact same cuts.

Preferences