Remove silence
from video.
Less waiting between the words. Find the pauses, choose your pace and keep every moment that matters. Save a tighter video with picture and sound together.
Find the gaps. Keep the rhythm. Save your next take.
Video silence remover is in your dock. Everything you started is still there.
Let the words flow.
Shorten the pauses. Keep the moments that matter.
Your next take, with less waiting.
Drop a video with speech. You decide which pauses go.
MP4MOVWEBMMKVM4V
Preview skips the marked pauses. Saving makes one continuous video.
Click or drag to explore. Cut times refer to the original recording.
Find your pace
Start with a feel. Fine-tune below.
Natural keeps a little space between thoughts.
Give words some room
An edit should still sound like you.
Speech detection
Check quiet words after changing this. Restored pauses reset when the detected speech changes.
Keep the good take
Review it once before you save.
MP4 · original shape · 30 fps
Picture and sound are cut together. Both are re-encoded. Extra tracks and embedded subtitles are left out.
Your video, detected speech and cut decisions stay in this tab.Only your preferred output format saves privately.Saving details
From a long pause to the next thought.
Open a spoken recording
Choose a video or try the example. The tool reads its soundtrack and finds speech on your device.
Make the pace feel right
Choose Relaxed, Natural or Tight. Listen around each pause. Untick any cut that removes a moment you need.
Save the edited video
Compare Edited with Original. Press Save video to make one file with picture and sound cut together.
A guide to the pacing studio
The tool proposes cuts. You decide which ones belong in the finished recording. Start with Natural and listen before making the edit tighter.
Open a video and compare the versions
Choose a video, drop it anywhere on this page or use Try an example. Speech detection starts after the soundtrack opens. The first run fetches the speech model from floi. Your recording stays on your device.
| Control | What it does |
|---|---|
| Edited / Original | Preview with proposed pauses skipped, or play the untouched timing. The saved file always uses Edited. |
| Play, Preview position and Sound on/off | Review the recording. Muting the preview does not mute the saved file. Preview seeks can briefly hesitate at a cut. |
| The whole recording | Shows source time, the waveform and every proposed removal. Click or drag to move. A cut is skipped in Edited. |
| Open another / Clear | Replace the source or release it and return to the empty studio. Cancel an active operation first. |
| Narrow, Wide or Full | At the top right of the tool. Open the studio at a comfortable size. Escape or Close returns to the page. |
Find your pace
Relaxed leaves more breathing room. Natural shortens obvious waits. Tight brings spoken phrases closer together. A preset sets the three values below; you can change each value afterwards.
- Only cut pauses longer than protects short pauses from automatic removal. It measures the gap between detected phrases.
- Protect each side of speech keeps time before and after every detected phrase. Raise it if a quiet opening or ending feels clipped.
- Keep extra time per pause adds more quiet time between phrases, split evenly around the cut. Protection and this extra time add together.
- Trim quiet start and end also proposes sufficiently long gaps at the edges. Untick an edge cut if the opening or ending matters.
Review the pauses
Every row shows the source range that will go and its duration. A checked box means Cut. Untick it to keep the whole pause. Listen plays the join with about 1.2 seconds of context on each side.
Cut all selects every proposed removal. Keep all restores every pause. Undo reverses your last review decision, up to 25 decisions. Pacing changes keep restored pauses; they do not change your source file.
Protect quiet words
Detection balance changes how the model’s speech probabilities become regions. Protect quieter speech tends to retain more voice. Make more cuts requires stronger speech evidence and can remove quiet words.
Changing this setting resets restored pauses because the boundaries can move. Find speech again retries the local model if loading or analysis failed. If it finds no speech, the full recording stays.
Save picture and sound together
File format offers MP4 or WebM when your browser can encode both tracks. The output note shows its dimensions and 30 fps timeline. The long edge is capped at 1920 pixels without enlarging the source.
| Action | Result |
|---|---|
| Save video | Makes one edited file named with pauses-cut. Repeating Save reuses the finished file until an edit changes. |
| Cancel | Stops opening, speech detection or saving. Cancelling detection or export keeps the opened recording. An unfinished result is not saved. |
| Open in / Share | Prepares the same edited file for another floi tool or a sharing destination. These controls live in the file bar. |
| Ctrl / ⌘ + O | Opens the shared file picker. |
| Ctrl / ⌘ + S | Presses the available Save video control. |
Know when a pause should stay
Keep pauses that show an action, give an answer time to land or let a viewer read the screen. Speech detection does not understand those purposes. Music and noise can confuse it, and silence removal does not repair noise beneath speech.
This studio takes one SDR video up to 300 MB and five minutes. The source picture must be at most 8.8 megapixels. Mono and stereo sound are supported. These are application limits, and a device can run out of memory sooner.
For audio-only work, use the audio silence remover. For a single time range, use the video trimmer. Add fresh captions to the edited timing with the subtitle generator.
Frequently asked questions
Can I automatically remove silence from a video for free?
Choose a video and the tool finds speech locally. Review the proposed cuts, choose your pacing and save an MP4 or WebM. There is no account requirement, watermark or daily quota. The application accepts one clip up to 300 MB and five minutes.
Will it remove filler words, mistakes or bad takes?
No. Speech detection finds where a voice is present. It does not understand the words or decide which performance is best. An “um” can count as speech and stay in the video. Use a transcript editor or make those cuts manually.
Does it remove background noise from a video?
No. It shortens gaps between detected spoken phrases. Noise beneath retained speech remains. Music, overlapping voices and very quiet speech can confuse detection, so listen to the proposed edits before saving.
How do I keep a dramatic pause or a silent demonstration?
Untick that pause in Review the pauses. It stays in the result even when you change the pacing values. Undo reverses a review decision. Changing Detection balance can move speech boundaries, so it resets restored pauses.
Does removing pauses reduce video quality?
Picture and sound are re-encoded, so this is not a lossless edit. Output uses even dimensions within a 1920-pixel long edge, at 30 fps. Rounding can change its proportions slightly. Your source stays unchanged, and HDR is refused.
Are my video and cut decisions saved to my account?
No. The recording, waveform, detected speech, previews and cut decisions stay in tab memory. Reloading loses that work. Only your preferred output format can save automatically to a signed-in account. The speech model loads from floi’s model host after interaction.
What happens to existing subtitles?
Embedded subtitle tracks are not copied. Burned-in words are part of the picture and stay visible where their frames survive. An external subtitle file still uses the original timing. Generate or retime captions after cutting the pauses.
Something missing? Request a feature