Drop it in
Interviews, voice notes or an entire lecture. Choose a file and its language.
Audio to Text Transcription
A transcript you can edit. Words you can find. Choose a recording to get started.
Drop your audio here
Audio or video files
Choose several files to transcribe them one after another.
A transcript you can work with
00:04
Click a word.
Hear the moment.
Edit the text. Label the speakers.
Take your transcript anywhere.
Several files
Batch
Free: up to 30 minutes per file, one file at a time. With credits: up to 3 hours per file, two at a time, with speaker labels. Finished transcripts are saved in Saved transcripts on this device; keep this tab open until files finish.
This browser cannot read the duration. Enter the recording length; the processor checks it before transcription.
Get creditsNo signup, and your file stays here.
Free up to 30 minutes, 2 files a day. Credits add longer files (up to 3 hours), speaker labels and batches.
Uploading
Live transcript
Words appear as each 30-second section finishes. The editor opens when the whole recording is done.
Transcript ready
Your transcript
0:000:00
Click a word to hear it. Choose Edit to make a correction.
Edits save on this device.
Find & replace
Replace text across the whole transcript, including passages outside the current search. Undo restores the whole change.
Enter the text to find.
Take it with you
Your edits are included in every export.
Plain text (.txt).
Review names, numbers and speaker labels before sharing. Automatic transcripts can make mistakes.
Translate
Machine translation, included with your credits. Review it before use.
Uses your wording when you choose Translate. Later edits do not update an existing translation.
For a translated file, choose Text, Word, SRT, VTT, CSV or JSON above. Captioned video uses the original language.
Need to know who said what? Transcribe again with speaker labels, using credits. This transcript stays in Saved transcripts.
Make it yours
Listen, search and fix a word. Name each speaker. Your edits save as you go.
Put it to work
Save a document, copy a quote or download captions with their timestamps.
From a recording to a useful document
Choose a file from your device, select its language or let StemVocal detect it, and start transcription. You can try two free jobs each day without an account, with recordings up to 30 minutes each. These free jobs are shared with the other StemVocal tools and reset at midnight UTC. For longer recordings or more jobs, choose a file pack, the monthly Transcription plan or Unlimited. One credit covers each started 30 minutes, up to three hours per file. Unlimited includes recordings up to three hours without spending credits. Files have a 1 GB limit. Transcription uses Whisper large-v3-turbo; a free job may use the smaller Whisper base instead. Neither guarantees a word-perfect transcript.
Audio transcription turns speech into text you can search, quote and edit. Upload MP3, M4A, WAV, FLAC, OGG, AAC, Opus, AIFF or WMA files from interviews, voice memos, podcasts or lectures. Choosing a file opens a local preview. Video recordings work too: an MP4, MOV, WebM or MKV, such as a meeting recording from Zoom, Teams or Meet or a clip from your phone, is transcribed from its audio track. AIFF, WMA, AVI and MPEG files may need credits. There is no app to install, and you do not have to convert every recording to MP3 first.
Listen once, correct in context
Click a word to hear that moment, slow playback down to check a name, and search the complete transcript for a phrase. Choose how many people speak if you want suggested speaker labels, which use credits, then rename them in the editor. Labels are estimates, not identity recognition; short turns, overlapping voices and similar voices need a careful review.
The editor saves changes on this device. Uploaded recordings and the original transcript remain accessible for 24 hours, so download anything you need to keep. A text file is useful for notes, Word for further editing, and CSV or JSON for timestamped data. SRT and VTT exports turn your transcript into subtitle cues. Print / PDF uses your browser’s print dialog to produce a readable document. Text, Word and print can include a timestamp for each turn. A paid transcript can also be translated into English, or from English into 15 other languages, at no extra cost. Choose a text or subtitle format to download it. Translation uses your wording when you choose Translate; later edits do not change an existing translation. Review machine translation before use.
For a recurring name or term, open Find & replace. Check the match count, choose whether to match case or whole words, and replace it across the transcript. One Undo restores the whole replacement.
Choose a transcript format
Before downloading, choose a format for the next step:
| What you need | Export | What to check |
|---|---|---|
| Notes you can copy into another app | TXT | Speaker names and corrected wording |
| A document for an editor or colleague | Word (DOCX) | Paragraphs and speaker labels |
| A transcript to print or save as a PDF | Print / PDF | Page layout in your browser’s print preview |
| Captions for an audio player or an accompanying video | SRT or VTT | Cue timing against the original recording |
| Timestamped turns for a spreadsheet | CSV | Start/end seconds and speaker labels |
The downloadable transcript examples show the plain-text, subtitle and CSV formats before you choose a file.
More tools for your recording
For screen recordings and camera footage, use StemVocal’s video studio, with a video player and caption preview. For music rather than speech, separate vocals and instruments with the music tool. Sign in with the same email on each site to use one shared wallet.
Return to saved work
Open Saved transcripts to pick up an earlier transcript, correct it or download it again. Your saved work stays in this browser, on this site. Reconnect the same original recording to listen while editing; reopening text uses no credit. You can delete each local copy separately. Download important work, since browser storage can be cleared or evicted.
Get a cleaner transcript
A clear recording makes the biggest difference. Keep the microphone near the speaker, reduce background music, and select the spoken language when automatic detection is uncertain. Transcription can miss words or introduce mistakes, especially in noisy or quiet passages. When people talk over each other, the transcript usually keeps one voice. English, Spanish, French, German, Portuguese, Italian, Dutch, Russian, Ukrainian, Polish, Turkish, Indonesian, Vietnamese, Japanese, Korean and Mandarin Chinese were measured on clear read speech; the other languages in the menu work with more variable accuracy. Review names, figures and important quotations against the original before using the text. This is an editing tool, not a certified transcription service.
See what your transcript downloads contain
A text download preserves the corrected words and speaker labels. For timestamps you can sort or analyze in a spreadsheet, choose CSV. This written example demonstrates the export format; it is not a transcription accuracy sample.
Interviewer: What should we check before publishing?
Speaker: Check names and numbers against the recording. Then save the corrected transcript.
Example TXT · Example SRT · Example VTT · Example CSV
TXT is for reading, CSV keeps one row per speaker turn, and SRT/VTT divide text into timed subtitle cues. Export files contain your edits; keep the source recording separately.
Good to know
Is transcription free?
Yes. You get 2 free StemVocal jobs each UTC day, shared across audio transcription, video transcription and music separation, for recordings up to 30 minutes each. Extra jobs, longer files and speaker labels use credits: a credit pack, the $9/month Transcription plan (up to 30 hours) or Unlimited.
How many credits does a recording use?
One credit per started 30 minutes: 20 minutes uses one credit, 45 minutes uses two, and 3 hours uses six. You choose the paid action before credits are reserved. Failed jobs return their reserved credits to the original allowance. Unlimited includes these jobs without spending credits.
Can I close the tab?
Keep it open until the upload finishes; then you can return on the same device within 24 hours. An interrupted upload can resume its completed parts. If a free transcription runs in your browser, keep the tab open until it finishes, because closing it stops processing. Finished transcripts and edits save on this device; open Saved transcripts to return to earlier work.
What happens to my files?
Your recording is uploaded to private storage for transcription, with 24-hour access and deletion normally within another 24 hours. You can delete it sooner from the editor. When a free transcription runs in your browser instead, the recording and text stay on your device. Saved transcripts and edits remain in this browser until you delete them, clear site data or the browser evicts them.