Transcribe video to text with every word timed to the footage.

Cutwords is an AI transcription tool for video and audio. Upload a recording and get a transcript with every word timed, in 25 languages, ready for you to name and assign speakers. Fix words by typing, then edit the video from the text or download the transcript and subtitles.

Start editing free Try it on this page

Free plan included. Paid plans are billed at checkout.

25 languages: plus Auto-detect, so you don't have to pick one
Word-level timing: click any word and the video jumps to it
0 AI credits: to transcribe and label speakers, on every plan including Free
SRT and TXT: plus edit lists for Premiere Pro, Final Cut Pro and Resolve

Accurate to the word, with every word linked to its moment

Every word in your video gets its own timestamp from a large speech-recognition model, so clicking a word jumps the video to that moment. If a word was misheard, change it the way you'd fix a typo. Only the text changes; the audio and your cut stay exactly where they were.

Time every wordEach word knows where it sits in the recording
Jump to a wordClick it and playback starts there
Edit a wordF2Correct it in place; the timing stays put
Fix a name everywhereSave the spelling once in the glossary

Words light up in step with the footage, a click jumps to "pricing," and "lunch" is corrected to "launch."

Know who said what — name and assign speakers

Each paragraph of your video's transcript can be assigned to a speaker. Name someone once and every line they're assigned updates, along with your exports, captions and lower thirds. Move a paragraph to a different person in two clicks.

Label speakersAssign a name to each paragraph
Name a speakerName, title and photo, set once for every line
Reassign a paragraphMove a line to the right person
Filter by speakerRead only what one person said

"Speaker 2" becomes Jon Okafor in every paragraph at once, then one line moves to Maya.

Start editing when the upload finishes, not days later

Your video starts transcribing on its own as soon as the upload finishes, and the project shows each step as it happens. When it's done the transcript opens ready to read, search and edit, with filler words already found. There's no human transcriber in the loop and nothing to wait for by email.

UploadingProgress in percent while the file goes up
TranscribingStarts by itself when the upload finishes
Filler words foundMarked before you open the transcript
Transcript readyOpen it and start editing

A 48-minute interview goes from upload to an open transcript, step by step.

Filler words and long pauses, flagged as you read

Every um, uh and you know in your video is marked in the transcript, and each long pause shows its length. Filter to just the filler words or the pauses, then remove them in one go with /fillers or /pauses. Finding them uses no AI credits.

Mark filler wordsum, uh, you know and the rest of the list
Time long pausesEach one tagged with its length in seconds
Filter the transcriptEverything, Filler words, Pauses or one speaker
Add your own filler words (Soon)Words you say that aren't on the list

The filter isolates the filler words, then /fillers removes them and the runtime drops.

Spell names and terms your way, once

Fix a misheard name in your video's transcript everywhere it appears, in one step. Save the heard spelling and the one you want to your workspace glossary, then apply it to every match. The glossary is shared by your workspace, so the same correction is ready in every project your team opens.

Save a spellingHeard "Priya Shatty," preferred "Priya Shetty"
Locate each matchJump to it before you change it
Apply the fixTo the matches you select, all at once
Undo it⌘ZText only; audio and cut timing stay unchanged

A name saved in the Workspace glossary, then applied to its three matches in the transcript.

Transcribe 25 languages, with the language detected for you

Your video is transcribed in the language it was recorded in, picked from 25 or left on Auto-detect. An interview in Spanish, a lecture in German and a podcast in Japanese all get the same word-timed transcript, ready for you to name and assign speakers.

Auto-detectThe default; Cutwords works out the language
EuropeanEnglish, Spanish, French, German, Italian, Portuguese, Dutch, Polish, Swedish, Danish, Norwegian, Finnish, Czech, Greek
And moreRussian, Ukrainian, Turkish, Arabic, Hindi, Indonesian, Vietnamese, Thai, Japanese, Korean, Chinese

With Auto-detect on, a German lecture, a Japanese podcast and a Spanish interview each come back in their own language.

Turn the transcript into captions or an SRT file

Your video gets captions from the words you just checked, animated in any of 39 styles and kept in sync when you cut. Burn them into the export, or download an SRT subtitle file for YouTube, Vimeo or your own player.

Style captions39 styles, from Minimal to Karaoke Bar
Burn them inRendered frame by frame into the MP4
Download subtitlesAn .srt file, timed to your edit

The same line plays in three caption styles, then downloads as an SRT.

Download the transcript, or take the cut to your editor

Your edited video's words leave Cutwords as plain text, an SRT, or an edit list for Premiere Pro, Final Cut Pro or DaVinci Resolve. Everything you removed in the transcript stays removed in the file.

Transcript.txt with the words of your edit
Subtitles.srt timed to your edit
Edit listsAdobe Premiere, FCP, Resolve and EDL
Draft.json with the whole project and its timed transcript

The Export menu downloads the subtitles, then the transcript, with a preview of each file.

Next: translated transcripts, and misspoken words fixed in the speaker's voice

Soon your video's transcript can be translated into another language for subtitles. And when someone says the wrong word, you'll type the right one and hear it re-spoken in their own voice, without a reshoot.

Translate (Soon)Subtitles in another language from the same transcript
Fix a misspoken word (Soon)Type the correction; it's re-spoken in their voice

A preview of both: a line translated to Spanish, and "Tuesday" re-spoken as "Thursday."

How it works

Step 1
Upload a recording
Drop in a video or audio file. Pick the language, or leave it on Auto-detect.
Step 2
Read and fix
Name the speakers, correct a word by typing, and fix names everywhere with the glossary.
Step 3
Edit or export
Cut the video by editing the text, add captions, or download TXT, SRT or an edit list.

Who it's for

Podcasters: Get show notes, quotes and an edit from the same transcript.
Journalists: Search hours of interviews for the exact quote, with the moment it was said.
Educators: Turn recorded lessons into readable text and captions for every student.
Researchers: Transcribe interviews with each participant labeled, in their own language.
Video editors: Make selects by reading, then send the cut to Premiere Pro, Final Cut Pro or Resolve.
Marketing teams: Pull customer quotes from a webinar and caption the clips in one place.

Questions

How accurate is Cutwords transcription?

It uses a large speech-recognition model and gets the most right on clear audio: one voice close to the microphone, little background noise. Crosstalk, music under speech and distant microphones cause more mistakes. Every word is editable, and the glossary fixes a misheard name everywhere at once.

Which languages can it transcribe?

25: English, Spanish, French, German, Italian, Portuguese, Dutch, Polish, Russian, Ukrainian, Turkish, Arabic, Hindi, Indonesian, Vietnamese, Thai, Japanese, Korean, Chinese, Swedish, Danish, Norwegian, Finnish, Czech and Greek. Leave it on Auto-detect and Cutwords works out the language for you.

Does it tell me who's speaking?

You assign each paragraph to a speaker — Speaker 1, Speaker 2, or a real name. Rename a speaker once and every line, caption and lower third uses the new name, and you can reassign any paragraph to a different speaker.

Can I get subtitles?

Yes. Download an SRT file timed to your edit, or burn animated captions into the video in any of 39 styles.

Can it translate the transcript?

Not yet. Translation is coming soon. Today you get the transcript in the language that was spoken.

Can I edit the video from the transcript?

Yes. Delete words and the matching video and audio leave the edit. That's the text-based editor, and it works on every transcript.

Does transcription use AI credits?

No. Transcribing and labeling speakers don't use AI credits on any plan, including Free.

Can I fix mistakes in the transcript?

Yes. Select a word and edit its text (F2), or use the workspace glossary to correct a name everywhere it appears. Only the text changes; the audio and your cuts stay where they were.

Is there a free plan?