Cut the silence, caption the rest.
Three tools. No login. FastAPI · faster-whisper · ffmpeg.
Cut silence from a video
ffmpeg silencedetect finds the gaps; we trim them and re-concat the rest. Audio and video stay in sync.
lower = stricter. -30 to -40 typical for speech.
gaps shorter than this are kept.
keep this much audio at each cut edge.
drop keep-segments shorter than this.
if total kept < this, skip the cut.
drop a video to start
queued
—
Transcribe + burn subtitles
faster-whisper transcribes audio → SRT, optionally burned into the video with ffmpeg subtitles filter.
faster-whisper int8 quantization. base ≈ 1 GB RAM.
78 ≈ TikTok lower-third · 50 = center
drop a video to start
queued
—
Burn your own SRT into a video
Upload a video and a .srt file — same styling options as the subtitle tab.
upload both files to start
queued
—
Cut silence + add subtitles in one click
The full TikTok/Reels pipeline. Silence gets cut first, then whisper transcribes the result, then subtitles get burned in. One upload, one download.
drop a video to start
queued
—