SPEECH TO TEXT · SPEAKER LABELS
Recordings in. Labelled transcripts out.
Upload a meeting, interview or podcast. Every turn comes back with a timestamp and the right speaker beside it, ready to fix up and export.
- About 200 minutes free each month
- No credit card
- MP3, WAV, M4A and more
- MayaThanks for coming in. Let's start simple. Why open a bakery in a town this small?
- DanielHonestly, because there wasn't a bakery here. People were driving across the valley for decent bread.
- MayaAnd when you opened, what surprised you most?
- DanielHow early everyone wakes up. We sold out before breakfast most Saturdays.
- MayaSo you started baking in the middle of the night?
- DanielBefore dawn every day. My partner still hasn't forgiven me.
- MayaWhat's next for you?
- DanielA small counter at the train station if the council says yes.
- TXT
- SRT
- VTT
A recorded SmileToAI run: text and timings as returned, speakers renamed in the studio. Recorded in English.
What comes back with every transcript
Speaker labels (diarization)
Each turn is tagged with who said it. Rename Speaker 1 to Maya once and every line updates; move a mislabelled turn to the right person.
A timestamp on every turn
Find the moment in the recording when someone said it. Handy for quotes, show notes and chapter markers.
Export to TXT, SRT or VTT
Plain text for documents, or subtitle files for your video editor, with long turns split into short cues that follow the speech.
Language detected for you
Leave the language on auto and it is detected from the audio. Or pick one yourself.
Made for recordings with more than one voice
Meetings
See who agreed to what without scrubbing back through the call.
Interviews
Your questions and their answers come back apart, so quotes are quick to find and check.
Podcasts
A turn-by-turn record of hosts and guests for show notes, quotes and chapters.
Research and focus groups
Every remark tied to the right participant, ready for your analysis.
How it works
- STEP 1
Upload a recording
MP3, WAV, M4A, AAC, OGG or FLAC, from your computer or your phone.
- STEP 2
Transcribe with speaker labels on
Labels are on by default. Switch them off for a single-voice recording.
- STEP 3
Review, fix and export
Rename speakers, correct a turn, then download TXT, SRT or VTT.
WHAT IT COSTS
A minute with speaker labels ≈ 25 Sparks
The free plan's 5,000 monthly Sparks cover about 200 minutes with speaker labels on.
Questions
Is speaker diarization included?
Yes. Speaker labels (diarization) are available on every transcript and on by default. Turn them off for a single-voice recording.
Which audio formats can I upload?
MP3, WAV, M4A, AAC, OGG and FLAC.
Can I fix a speaker that was labelled wrong?
Yes. Rename any speaker, or move a turn to the right person. Your corrections are saved with the transcript.
Can I make subtitles from it?
Export SRT or VTT. Long speaker turns are split into short subtitle cues timed to the speech, so the text never covers the whole video.
How much does it cost?
About 25 Sparks a minute with speaker labels on. The free plan's monthly Sparks cover about 200 minutes.
Does it work on my phone?
Yes. The same studio runs on the web, iOS and Android.
Guides & resources
Your next recording, transcribed.
About 200 minutes free each month. No credit card.
Start freeWeb · iOS · Android