MP4 to text, without the MP3 detour
Drop an MP4 in and get its words out. The audio track is pulled from the container for you, so there is no export-to-MP3 step in between — and no quality lost to a second round of compression.
Drop your MP4 here
MP4, MOV, AVI, MKV, WebM — up to 120 minutes
MP4 is the file you almost always end up with
Phones, screen recorders, meeting software, mirrorless cameras — ask almost any device to produce a video and what lands on your drive is an MP4. It won that position by being the one container that plays everywhere without anyone having to think about it.
That ubiquity is also why so many people need to transcribe MP4 to text at some point. The recording was never made with transcription in mind; it was just the default output of whatever captured the moment.
Which leads to the thing worth understanding before you convert MP4 to text: an MP4 is a container, not a codec. Inside it sit two separate streams — usually H.264 or H.265 for the picture and AAC for the sound. Everything below follows from that one fact.
The picture is 95% of the bytes
In a typical MP4 the video stream dwarfs the audio one. Everything you want from a transcription lives in the few percent that carry sound.
Only the audio gets read
The video stream is discarded before speech recognition starts. A 4K file and a 480p file of the same length produce the same transcript for the same price.
You rarely chose the format
MP4 is what the device handed you. That also means the audio inside was recorded for watching, not for reading — which is what shapes accuracy.
Size is the real constraint
Minutes rarely stop you first. On phone footage the file hits the size ceiling long before the length limit does.
What converting MP4 to text actually involves
Converting MP4 to text means demuxing the file — separating its audio stream from its video stream — and running speech recognition on the audio alone. The picture is discarded before transcription begins, which is why resolution changes neither the price nor the result.
Four things follow from that definition, and together they answer most of what people ask before uploading their first file.
- Demuxing, not converting Nothing is re-encoded on the way in. The existing audio stream is lifted out of the container intact, so no quality is lost at this stage.
- No MP3 export needed Older guides tell you to extract an MP3 first, because early tools only accepted audio. That step is now redundant, and re-compressing to MP3 can only make the audio worse.
- Timestamps come from the audio Each line is stamped with when it was spoken, which is what lets the same transcript become an SRT subtitle file without extra work.
- Already have audio? If your file is an MP3, WAV or M4A rather than a video, the audio tools handle it directly — same engine, same credits, one less step.
Transcribe MP4 to text in four steps
The whole thing runs in the browser. Nothing to install, and no need to prepare the file beforehand.
Upload the MP4 as it is
Drag the file in straight from your drive. It uploads directly to storage from your browser, so a large recording is not held up by any intermediate size cap along the way.
The audio track is pulled out
The container is opened, the sound is separated from the picture, and only the sound goes forward. This happens automatically — it is not a step you perform.
Speech recognition runs
The spoken language is detected across 100+ languages and the transcript comes back split into timestamped segments you can read and correct.
Export what you need
Plain text for documents, or SRT and VTT if the words are going back onto the video as subtitles. One run, as many formats as you want.
A ten-minute MP4 typically comes back in under a minute. Longer files are split internally and stitched back together, so a two-hour recording behaves the same as a two-minute one.
The recordings people actually need transcribed
Almost none of these were filmed with a transcript in mind. That is the normal case, not the exception.
Meeting recordings
Exported from conferencing software as MP4. Usually the cleanest audio you will get, because the sound came down the channel rather than through the air.
Screen recordings
Tutorials, walkthroughs, bug reports. System audio plus a microphone, and long enough that scrubbing back through them is painful.
Lectures and training
An hour or more of one person talking — the format where a searchable transcript saves the most time per file.
Interviews on a phone
Shot on whatever was in someone's pocket. High bitrate, large file, and audio recorded from further away than anyone intended.
Video podcasts
The episode is already made; the show notes are not. A transcript turns the second job into an edit rather than a re-listen.
Event and conference footage
Talks worth quoting, buried in files nobody will scrub through. Text is what makes them citable.
45-minute webinar Comes back as roughly 6,000 words — enough raw material for an article, a newsletter and a handful of clips.
2-hour lecture Around 200MB at 480p, so it clears both limits comfortably. Split internally, returned as one transcript.
10-minute phone interview Can easily be 1.2GB at 1080p60. Well within the size ceiling, but a reminder that minutes and megabytes are not the same measure.
30-second clip Costs 250 credits — the billing follows the length of the recording, so short files stay cheap regardless of resolution.
What goes in and what comes out
MP4 and its relatives
MOV, AVI, MKV and WebM all work the same way — they are containers too, and the audio is lifted out of each of them identically.
TXT for reading
Plain text with or without timecodes. This is the one for notes, search, quoting, or pasting into a draft.
SRT and VTT for subtitles
Timed files that drop into YouTube, Premiere, Final Cut or any player. Name the SRT to match the MP4 and most players load it automatically.
100+ languages
Detected automatically. A recording that switches language partway through is transcribed in each language as it was actually spoken.
Subtitles are not burned into the picture — the SRT stays a separate file, so you can correct a name or a term without re-rendering the video.
Getting a clean transcript out of an MP4
The audio in an MP4 was almost always recorded before anyone thought about transcription. You cannot go back and re-record it, but knowing what hurts accuracy tells you what to fix and what to accept.
- Camera-mic distance Phone and camera microphones sit metres from the speaker and pick up the room along with the voice. This is the single biggest reason phone-shot MP4s transcribe worse than screen recordings.
- Screen recordings win When the audio comes from the system rather than through the air, there is no room, no distance and no handling noise. These files consistently produce the best results.
- Background music Music sitting under dialogue confuses recognition far more than steady noise like an air conditioner. Strip it before transcribing rather than after.
- Several people at once Crosstalk is the most common cause of garbled passages. Nothing can be done retroactively, so expect to clean those stretches by hand.
- Names and jargon Product names and technical terms are where even a 97% transcript shows its seams. A find-and-replace pass on the exported text fixes it in under a minute.
How long can your MP4 actually be?
Paid plans allow 120 minutes or 2GB per file, whichever comes first — and which one you hit depends entirely on how the recording was made. Bitrate decides it, not duration. These figures are estimates from typical bitrates; your own file may differ.
| Source | Typical bitrate | 2GB holds | 120 min weighs | Limit you hit first |
|---|---|---|---|---|
| Phone, 4K | 45 Mbps | ~6 minutes | ~40 GB | Size |
| Phone, 1080p60 | 25 Mbps | ~11 minutes | ~22 GB | Size |
| Phone, 1080p30 | 17 Mbps | ~16 minutes | ~15 GB | Size |
| Screen recording, 1080p | 8 Mbps | ~33 minutes | ~7 GB | Size |
| Meeting export, 720p | 2 Mbps | ~2.2 hours | ~1.8 GB | Duration |
| Lecture, 480p | 1 Mbps | ~4.4 hours | ~0.9 GB | Duration |
If your file is too large: re-encode it at a lower bitrate before uploading. Dropping phone footage from 1080p60 to 720p shrinks it by roughly ten times and changes the transcript not at all, because the audio track is untouched by video re-encoding. Splitting the file also works, though you will then be stitching two transcripts together by hand. What does not help is converting to MP3 first — same audio, extra step, and one more round of compression.
What the text unlocks
Every tool below draws on the same credit balance as your transcription — nothing here needs a second subscription.
Text to Speech
Turn the edited transcript back into narration in a different voice.
Open the MP4 to text converterVoice Cloning
Narrate it in your own voice, trained from a short sample.
Open the MP4 to text converterVoice Isolator
Strip room noise and background music — run it before transcribing, not after.
Open the MP4 to text converterAI Podcast Generator
Turn a long transcript into a multi-speaker episode.
Open the MP4 to text converterQuestions about MP4 transcription
Convert your first MP4 to text free
New accounts include about ten minutes of transcription. Enough to run a real recording through and judge the result on your own footage rather than a demo clip.
Convert MP4 to text