Start free

Transcribe WhatsApp voice messages to text.OPUS and OGG accepted.

Export the WhatsApp voice note from any chat, drop the .opus file in as-is. Text back in seconds — no MP3 conversion, no 'unsupported format' errors, 99 languages including code-switching.

Drop a file, or pick one

MP3 · WAV · M4A · MP4 · MOV · MKV · OGG · OPUS · FLAC · WEBM — up to 100 MB anonymously

Paste a link, we’ll fetch the audio

YouTube · TikTok · Vimeo · Twitter · SoundCloud · Spotify · 50+ more

Record straight from your browser

Sign up takes 30 seconds — recording opens right after, in the dashboard.

No card required~90s per 60-min fileSRT · VTT · DOCX · TXTFiles auto-deleted in 24h

↓ Watch what comes out

Voice note in. Readable text out.

WhatsApp encodes voice notes as 16 kHz mono OPUS at roughly 24 kbps — compact, but not what generic transcription tools expect. We decode OPUS natively and skip the conversion dance.

WhatsApp voice exportREC PTT-20260512-WA0007.opus · 1:24
auto-detected es-MX → en16 kHz mono · OPUS 24 kbps
~90s
Transcript · streaming91% accuracy
S1

Hey, quick one — I'm running late for dinner, traffic on the bridge is insane.

S1

Can you tell Priya to start without me? I'll be there by 8:30 at the latest.

S1

Y dile que pida lo de siempre, the pad thai with extra peanuts.

S1

Okay, gotta go, calling you when I park.

91% on 16 kHz OPUSTXT · DOCX · SRT · JSON

↓ This is the dashboard

This is what loads when the job finishes.

Same layout as the real dashboard — Summary, full Transcript, Speakers tab, Exports. Key points and action items extracted automatically. Auto-tags on every job.

Try it on your own file — it's free

Three real options · honest comparison

WhatsApp built-in. Generic transcribers. Or us.

WhatsApp now transcribes voice notes inside the app on recent iOS and Android — but only in a handful of languages and only on-device. Generic transcribers refuse the .opus file. We take the export and run it.

Option 01

WhatsApp built-in

On-device transcription inside the chat bubble. Limited languages, no export.

RequiresiOS 16+ / Android 11+
Languages~9 supported
Code-switchingPicks one, drops the rest
ExportCopy-paste only
Bulk / archiveOne at a time
CostFree
Best forReading one English or Spanish voice note in the moment without leaving WhatsApp.
Option 02

Transcription.Solutions

Drop the .opus or .ogg as exported. Multi-language, exportable, batch-friendly.

RequiresJust the exported file
Languages99, auto-detected
Code-switchingHandled mid-sentence
ExportTXT · DOCX · SRT · JSON
Bulk / archiveDrag a folder of .opus
Cost · per min$0.03
Best forAnyone exporting voice notes from chats — journalists, researchers, lawyers reviewing client messages, bilingual families.
Option 03

Otter / Rev / generic

Won't accept .opus. Convert to MP3 first, lose quality, lose minutes.

RequiresConvert OPUS → MP3/WAV
LanguagesEN-tuned, few others
Code-switchingOne language per file
ExportAfter conversion + upload
Bulk / archiveNo batch OPUS in
Cost$8–17/mo + conversion time
Best forPeople who already converted the file and only need English.

WhatsApp built-in transcription availability and language list accurate as of 2026. Competitor OPUS support checked on free tiers.

Specific to WhatsApp

Three things that break generic transcribers on WhatsApp exports. generic transcribers.

Most tools were built for meeting MP4s, not phone voice notes. The codec, the languages, and the burst length all trip them up.

What goes wrong

  1. 1OPUS upload rejected. Otter, Rev, Trint expect MP3/WAV/M4A. You convert in a browser tool, get a 96 kbps MP3 from a 24 kbps source — quality stays bad, accuracy doesn't improve.
  2. 2Code-switching gets butchered. A note that flips between Spanish and English mid-sentence — common in bilingual families and WhatsApp's biggest markets — gets transcribed as the wrong language for one of the halves.
  3. 3Sub-5-second clips. Quick 'yes, on my way' replies are too short for VAD models to lock onto. Generic engines return empty strings or hallucinated text.

What to flip here

  1. 1Drop the .opus or .ogg file as-exported from WhatsApp. We decode OPUS directly — no re-encode, no quality loss from a second pass.
  2. 2Leave language detection on auto and turn on the multi-lingual flag in the job form. The recognizer tracks language per-segment instead of per-file.
  3. 3For short bursts, batch-upload a folder of voice notes. We pad and process them together so VAD has enough signal to anchor on.

Recommended job settings for WhatsApp voice notes

Drop a .opus and these flip on by default. Override per-job from the form.

Input format
OPUS · OGG · native (no convert)
Language
Auto-detect · multi-lingual on
Diarization
Off (single sender)
Filler words
Removed by default
Punctuation
Smart casing + sentence breaks
Export
TXT · DOCX · timestamped JSON

Accuracy · real-world numbers

91% on a quiet voice note. Drops fastest with wind and driving.

WhatsApp's 16 kHz mono OPUS caps what's recoverable — no stereo, no high frequencies. Numbers below are from real customer voice notes (de-identified), not synthetic test sets. The recording environment matters more than the codec here.

93%
Indoor, single speaker, one language

Quiet room, phone held normally. Native speaker, no code-switching. This is the realistic ceiling for OPUS at 24 kbps.

89%
Indoor with ambient noise

Kitchen, café, kids in the background. Filler words and overlaps from background voices may sneak in but rarely change meaning.

84%
Walking outdoors / on transit

Footsteps, traffic hum, occasional wind. Numbers and proper nouns are the first to slip. Expect a 1-minute proofread per 5-minute note.

76%
Driving with windows down

Wind buffeting, road noise above the voice. Words usable in chunks, full sentences less reliable. Worst case in our WhatsApp dataset.

Common questions

8 things people ask about WhatsApp voice transcription.

01How do I export a voice note from WhatsApp?+
Long-press the voice message → Forward → tap the share icon → save to Files (iOS) or share to Drive (Android). The exported file is a .opus or .ogg — drop that directly into our uploader, no conversion needed.
02Do I need to convert .opus to MP3 first?+
No. We decode OPUS natively, the same codec WhatsApp uses. Converting first actually hurts — you're re-encoding a 24 kbps source into MP3 and adding artifacts without recovering any quality.
03Does this work with Telegram, Signal, or Messenger voice notes?+
Yes. Telegram and Signal use the same OGG/OPUS format at similar bitrates. Messenger uses AAC in an .m4a container — also supported. Drop whatever the app exports.
04WhatsApp already transcribes voice notes — why use this?+
WhatsApp's built-in transcription is on-device, limited to about nine languages, and only readable inside the chat bubble. We give you exportable text, 99 languages, code-switching support, and batch processing for archives.
05Can it handle voice notes that switch between two languages?+
Yes — turn on the multi-lingual flag in the job form. We detect language per segment, not per file, so a sentence that starts in Hindi and ends in English transcribes correctly in both halves.
06What about voice notes forwarded between chats — any quality loss?+
None. WhatsApp forwards the original .opus file without re-encoding. A voice note forwarded five times sounds identical to the original — and transcribes the same.
07Group chat voice messages — do I get speaker labels?+
Each voice note is one sender, so diarization is off by default. If you upload several voice notes from different members, label them at upload (or by filename) and we keep the boundaries in the export.
08What happens to my voice files after transcription?+
Source audio is permanently deleted within 24 hours. Transcripts stay in your account until you delete them. We don't train models on customer audio — full policy on the security page.

Drop a WhatsApp voice note. See what comes out.

30 free minutes every month. No card. OPUS and OGG accepted as-is, 99 languages, all exports included.

Start free