Transcription Software with Speaker Identification (Speaker Labels)
Data checked:
Speaker identification (diarization) turns one wall of text into “who said what”. For meetings, interviews and focus groups it is usually the difference between a usable transcript and a useless one. Our data records three states per tool: labels on every plan, labels on paid plans only, and no labels, plus “not verified” where the vendor does not say. One caveat that applies to all of them: a label is usually “Speaker 1”, not a real name, and we have not tested how accurate any tool’s separation is.
Meeting tools that label speakers
| Tool | Speaker labels | Free plan | Paid from | Languages |
|---|---|---|---|---|
| Otter.ai Meeting notes, File transcription, Live transcription | Yes | 300 min/month, up to 30 min per conversation, 3 file imports in total | $16.99/mo $8.33/mo billed yearly (Pro) | ENJAZH 6 languages |
| Fireflies.ai Meeting notes, File transcription, Live transcription | Yes | Transcription marked unlimited, but only 400 min of storage per team and limited AI summaries; downloads need Pro | $18/mo $10/mo billed yearly (Pro) | ENJAZH 100+ languages |
| Notta Meeting notes, File transcription, Live transcription | Yes | 120 min/month but only the first 3 min of each recording, 50 file uploads/month; exporting transcripts needs Pro | $13.49/mo $8.17/mo billed yearly (Pro) | ENJAZH 58 languages |
| tl;dv Meeting notes | Yes | Unlimited recordings and transcripts, AI notes for 10 meetings, up to 5 uploads; recordings kept up to 3 months | €29/mo €18/mo billed yearly (Pro) | ENJAZH 30+ languages |
| Fellow Meeting notes, File transcription | Yes | 5 AI notes and 5 recordings per user for the lifetime of the account; 5 audio/video uploads per user lifetime. | $11/mo $7/mo billed yearly (Team) | ENJAZH 100 languages |
| Read AI Meeting notes, File transcription, Live transcription | Yes | 5 meeting transcripts per month. | $19.75/mo $15/mo billed yearly (Pro) | ENJAZH |
| jamie Meeting notes | Yes | 10 AI notes per month, with a 30-minute meeting limit. | $29/mo $24/mo billed yearly (Plus) | ENJAZH 99+ languages |
| Granola Meeting notes | Yes | Basic includes AI meeting notes and chat, with 30 days of meeting history. No monthly minute allowance was verified. | $14/mo (Business) | ENJAZH |
| Tactiq Meeting notes, File transcription, Live transcription | Yes | 10 transcripts and 5 AI credits per month; uploads count as meeting credits. | $12/mo $8/mo billed yearly (Pro) | ENJAZH |
| Bluedot Meeting notes | Yes | 5 meetings total for the lifetime of the free account, not 5 per month. | $18/mo $14/mo billed yearly (Basic) | ENJAZH 100+ languages |
Details worth knowing: Fellow confirms speaker identification in 100 languages, except Burmese, Khmer, Thai and Tibetan. Rimo Voice, a Japanese service, takes a different approach: you label a few lines by hand and AI assigns the rest, with voice-print-based separation on higher plans. OpenAI Speech to Text API can label speakers, but only through gpt-4o-transcribe-diarize, a model its guide marks deprecated until 26 February 2027.
File-upload services that label speakers
| Tool | Speaker labels | Paid from | Max length per file | Export |
|---|---|---|---|---|
| Sonix File transcription | Yes | $25/mo $22.92/mo billed yearly (Core) | Paid: Not verified | DOCX, PDF, TXT, SRT, VTT |
| Happy Scribe File transcription, Meeting notes | Yes | £17/mo £8.50/mo billed yearly (Basic) | Paid: Not verified | TXT, SRT, PDF, DOCX, VTT, HTML |
| TurboScribe File transcription | Yes | $20/mo $10/mo billed yearly (Unlimited) | Free: 30 min Paid: 10 h | DOCX, PDF, TXT, SRT, VTT, CSV |
| Transkriptor File transcription, Meeting notes, API | Yes | $9.99/mo (Lite) | Free: Not verified Paid: Not verified | TXT, SRT, DOCX |
| Trint File transcription, Live transcription | Yes | Not verified | Paid: Not verified | Not verified |
| Maestra File transcription, Live transcription, API | Yes | Not verified | Paid: Not verified | TXT, DOCX, PDF, JSON, SRT, VTT |
| Temi File transcription | Yes | Not verified | Paid: Not verified | DOCX, PDF, SRT, VTT |
Trint describes speaker recognition alongside collaborative editing; Maestra and Temi list speaker tags with subtitle exports; Scribie and Amberscript deliver speaker labels with their human-verified workflows. For Chinese and Japanese services, Feishu Minutes（飞书妙记） groups speakers by voice, and iFLYREC（讯飞听见） and Tencent Meeting Transcription（腾讯会议转写） list speaker separation.
Offline apps that label speakers
| Tool | Platforms | Speaker labels | Export | Languages |
|---|---|---|---|---|
| MacWhisper File transcription, Offline / on-device | Mac | Yes | TXT, SRT, VTT, DOCX, PDF, MD, HTML, CSV | ENJAZH 100 languages |
| Vibe File transcription, Offline / on-device | Mac, Windows, Linux | Yes | SRT, VTT, TXT, HTML, PDF, JSON, DOCX | ENJAZH |
| Buzz File transcription, Offline / on-device | Mac, Windows, Linux | Yes | TXT, SRT, VTT | ENJAZH 100 languages |
| noScribe File transcription, Offline / on-device | Windows, Mac, Linux | Yes | HTML, TXT, VTT | ENJAZH |
| aTrain File transcription, Offline / on-device | Windows, Linux, Mac | Yes | Not verified | ENJAZH 99 languages |
MacWhisper lists automatic speaker recognition without saying whether it needs the Pro licence; Vibe and Buzz include diarization free and open source; noScribe adds a transcript editor built for interviews; aTrain makes speaker detection optional. Since these run on your machine, labeled transcripts never leave it.
Where speaker labels cost extra
- AssemblyAI: diarization is a paid add-on at $0.02 per hour on top of $0.15 per hour for Universal-2.
- Good Tape: speaker labels come with Pro (€15 a month); the free entitlement was not verified.
- Texter: labels are a Premium feature (¥1,800 a month).
- superwhisper: labels need Pro.
No labels, or none we could verify
Aiko’s developer says speaker identification is not supported yet, and AudioPen records none. “Not verified” is not the same as “no”: for Rev, Fathom, Avoma, MeetGeek, Krisp (AI Meeting Notes), Riverside and VEED, the pages we checked simply do not say, so confirm with the vendor before paying.
For research interviews specifically, speaker labels interact with consent and export requirements; see transcription software for interviews. Every tool’s detail page links the vendor source for its speaker-label claim, and you can filter the full directory yourself.