Transcription with Speaker Identification: See Who Said What
Upload a conversation and get a transcript split by speaker, with a timestamp on every turn
No account? The first 30 minutes of any file are free. Create a free account for premium on your first file.
Upload a recording with two or more people and the transcript comes back split into turns, each labeled Speaker 1, Speaker 2 and so on, with the time it starts. Rename each speaker once and the name appears on every line and in every export.
Why label speakers here
Turns, not a wall of text
Each change of voice starts a new block with its speaker label and start time, so a two-hour conversation stays readable.
Rename once, used everywhere
Change Speaker 1 to Dr. Lee and the name follows into the editor, the Word file, SRT cues and VTT voice tags.
Fix a wrong label
If a line is given to the wrong person, move it to the right speaker in the editor before you export.
Exports keep the speakers
Word, TXT, PDF, CSV, SRT and VTT can carry speaker names, or leave them out when you want clean text.
How it works
Upload the conversation
Any audio or video file, or a link to a public recording. Speakers are detected for you.
Name the speakers
Open the transcript, listen to a line from each voice and give each speaker a name.
Export
Download with speaker names in the format you need, or copy the text.
Speaker labels: how to get clean turns
What a labeled transcript looks like
Every block in the transcript starts with a speaker label and a timestamp, followed by what that person said until the next voice comes in. Click any line to hear that moment, which makes it quick to check who is talking before you rename.
| Time | Speaker | Text |
|---|---|---|
| 00:00 | Speaker 1 | Thanks for making the time today. |
| 00:03 | Speaker 2 | Happy to. Where do you want to start? |
Get better speaker separation
- Record each person on their own microphone when you can, or sit everyone at a similar distance from one microphone.
- Ask people not to talk over each other during the parts you will quote.
- Keep the original recording: a compressed copy of a copy loses the voice detail that separates speakers.
- Very short turns, like yes or right, are the most likely to land on the wrong speaker. Check those first.
Formats that keep speaker names
Word and PDF put the speaker name above each block. CSV gives Start, End, Speaker and Text columns for a spreadsheet or coding tool. SRT puts the name at the start of each cue and VTT uses voice tags. TXT shows the name and time in each block header.
TXT downloads free on every file with a free account. Word, PDF, SRT, VTT, CSV and JSON come with your first file, the Weekly Pass, a minutes top-up or a paid plan.
Speaker diarization explainedHow to fix wrong speaker labelsTranscribing recordings with multiple speakers
What it costs
| You have | What you get |
|---|---|
| No account | The first 30 minutes of any file transcribed, readable on screen |
| A free account | Up to 30 files a day, up to 30 minutes each, TXT download on every file; your first file runs up to 60 minutes with every export format |
| Weekly Pass | $9.99 for 20 hours over 7 days, full-length files and every export format |
| Pro | $14.99/month or $99.99/year, full-length files, every export format and the AI tools |
Simple pricing
Free for files up to 30 minutes, no card needed. Paid plans start at $14.99.
Cancel anytime. See all plans
Frequently asked questions
The engine listens for changes of voice and groups each stretch of speech under a speaker label: Speaker 1, Speaker 2 and so on. It does not know who people are, so you rename the labels yourself in the editor.
Speaker labels come with our premium engine: on every file with a paid plan or pass, on files transcribed without an account, and on up to 2 hours a month with a free account. The first 30 minutes of any file are transcribed free with no signup.
Not yet. The number of speakers is detected from the audio. If two voices are merged or one voice is split, move the lines to the right speaker in the editor.
People talking over each other, very short replies, similar voices, and one microphone far from some speakers. A separate microphone per person, or a quiet room, gives cleaner turns.
They can. SRT cues can start with the speaker name, and VTT uses voice tags. You can turn speaker names off before you download.
Related tools
Interview Transcription
Upload the full interview or focus group, even 2 or 3 hours. Get each speaker labeled, with timestamps.
Meeting Transcription
Upload the whole meeting, even a 3-hour one. Get every word, who said what, and a summary.
Qualitative Research Transcription
Transcribe research interviews and focus groups with speaker labels, timestamps and exports ready for coding
Podcast Transcript Generator
Turn podcast episodes into searchable, shareable text
Audio to Word
Transcribe audio and download it as an editable DOCX file