Upload a video and get its spoken content back as accurate, timestamped text — with speaker labels where they can be detected. Download as TXT or SRT.
Drop your video here
MP4, MOV, WebM or MKV — up to 100MB
Uploading is free and needs no account. Free tier: 5 minutes per month. Running this tool needs an account, because it uses real processing time.
5 minutes per month
How It Works
MP4, MOV, WebM or an audio file. Only the soundtrack is read.
The audio is transcribed with timestamps, and separate speakers are labelled wherever the recording makes that possible.
Plain text for reading, or SRT with timecodes for subtitle work.
It extracts the audio from a video file, runs speech recognition over it, and returns what was said as written text. Unlike a YouTube transcript tool, which reads captions someone already made, this works on your own footage — files that have never been published and have no existing caption track.
Good transcription output is more than a wall of words. Timestamps let you jump back to any moment. Speaker labels let you follow a two-person conversation. Punctuation and casing turn a stream of recognised words into sentences a person can actually read.
The common uses: searching a long recording for a specific moment, turning an interview into a written article, pulling quotes from a podcast, meeting a compliance requirement for a written record, and feeding the text into anything that needs to understand the content rather than watch it.
Why Use This Tool
Jump from any line in the transcript back to the exact second it was said.
Separate speakers are identified where the recording makes it possible — the difference between a readable interview and an undifferentiated block.
With automatic detection, so you do not have to tell it what language the file is in.
Read it, edit it, or use it directly as a subtitle track.
Uploads are processed to produce your transcript and are not published anywhere.
Five minutes a month, free. No card, no countdown.
Upload the file here. The free tier covers five minutes of audio per month, which is enough for a few short clips or one long-form segment.
Typically high nineties on clear speech with a close microphone. Background music, overlapping speakers, heavy accents and room echo all reduce it.
Yes, where the audio supports it — distinct voices on a reasonably clean recording. Two people talking over each other on one microphone is the case that defeats it.
Different work. The YouTube tool reads a caption file that already exists. This one runs a speech recognition model over your audio, which has a real cost per minute.
Yes — MP3, M4A and WAV are accepted alongside video formats.
Five minutes of audio per month on the free tier, taken in one file or several.
Turning a recording into published short-form means transcribing it, finding the moments worth keeping, reframing them vertically, and captioning them. This page does the first. ShortsHero does all four, per clip, across the whole recording.
See the full pipelineFree to start · No credit card required
Keep Going
Generate subtitles automatically from any video. Get burned-in captions or a downloadable SRT file with accurate word-level timing — free, no signup required.
Paste any YouTube URL and get the full transcript with timestamps. Copy it, or download as TXT or SRT. Free, unlimited, no login required.
Edit SRT subtitle files in your browser. Fix wrong words, shift every timecode to resync drifted captions, stretch timing for frame-rate changes, and re-download.
Upload a podcast episode and get a vertical, captioned, speaker-tracked clip back automatically. Built for turning long conversations into TikToks, Reels and Shorts.