Start free

Transcription for clinicians.CME lectures, research interviews, medical conferences.

Drop a CME lecture recording, a consented research interview, or a conference panel. Get a speaker-labeled transcript with medical terminology handled — no BAA required, because this isn't built for PHI.

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

Conference recording in. Searchable transcript out.

Paste your drug list and acronyms into custom vocabulary before submitting and the recognizer biases toward them. Empagliflozin stays empagliflozin. NSTEMI stays NSTEMI.

CME panel · MP3REC 2 speakers · 38:12
auto-detected en-US44 kHz mono · 96 kbps
~90s
Transcript · streaming94% accuracy
S1

Dr. Patel, walk us through the GLP-1 titration schedule from your slide deck.

S2

We start semaglutide at 0.25 milligrams weekly for four weeks, then step up based on GI tolerance.

S1

And the cardiovascular outcomes from the SELECT trial — how did that change your practice?

S2

Twenty percent reduction in MACE. That's the number that moved us on secondary prevention.

94% on lavalier CME audioDOCX · SRT · TXT · 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

Human medical transcription. Generic AI. Or us.

Rev and similar human services will sign a BAA and transcribe patient encounters — but you pay per minute and wait days. Generic AI tools are cheap but stumble on drug names. We sit in the middle for the wide non-PHI use case: CME, research interviews, conferences.

Option 01

Rev human medical

Trained transcribers, HIPAA-aware, signs a BAA. Slow and expensive.

PHI / BAAYes (on request)
Turnaround12-48 hours
Medical vocabularyHuman-handled
DiarizationManual labels
Cost · per min~$1.50 (human)
ExportDOCX · TXT · SRT
Best forEncounters or PHI-containing audio where you legitimately need a covered entity workflow and can wait two days.
Option 02

Transcription.Solutions

Minutes, not days. Custom medical vocabulary. Non-PHI content only — audio deleted within 24h.

PHI / BAANot covered · non-PHI only
Turnaround~1× audio length
Medical vocabularyCustom hint list · 99 lang
DiarizationAuto · 2-8 speakers
Cost · per min$0.03
Audio retentionDeleted in 24h
Best forCME, grand rounds, consented research interviews, conference recordings — anything without identifiable patient audio.
Option 03

Otter / generic AI

Cheap and fast, but no medical biasing. Drug names and acronyms come back phonetic.

PHI / BAANo
TurnaroundFast
Medical vocabularyGeneric English model
DiarizationAcoustic, EN-tuned
Custom vocabLimited / paid tier
Cost~$17/user/mo
Best forCasual meetings where 'empagliflozin' becoming 'M paga floor zin' is funny rather than a problem.

Pricing and BAA availability accurate as of 2026. Rev's medical/BAA workflow is request-based, not the default product.

Specific to clinical content

Three things that break generic transcription on medical audio.

Most failures aren't acoustic — they're vocabulary. Flip these before you submit.

What goes wrong

  1. 1Drug names get phoneticized. Empagliflozin, dapagliflozin, semaglutide — generic models guess at syllables. You spend an hour with find-and-replace.
  2. 2Acronyms collapse. NSTEMI becomes 'in stemi'. CABG becomes 'cabbage'. ICD-10 gets read as a year.
  3. 3Panel Q&A merges speakers. Generic diarization expects two voices on a podcast, not four panelists plus audience questions.

What to flip here

  1. 1Paste your drug list, eponyms, and study names into custom vocabulary on the job form. We pass them to the recognizer as a bias, not a hard lookup.
  2. 2Add common acronyms with phonetic hints (CABG → 'cabbage' won't happen if CABG is in your vocab). Works for ICD-10 codes, trial names, gene symbols.
  3. 3Set speaker count to panel mode (4-8) and we tune diarization for handoffs every few seconds rather than long monologue blocks.

Recommended job settings for clinical content

Drop a CME or interview file and these flip on by default. Override per-job from the form.

Language model
en-US · medical vocabulary on
Custom vocabulary
Drug names, acronyms, trial IDs
Diarization
Panel mode · 2-8 speakers
Filler words
Removed by default
Summary
Key points + Q&A extraction
Export
DOCX · SRT · timestamped TXT

Accuracy · real-world numbers

94% on lavalier CME audio. Worst case is the back-of-the-room phone recording.

Medical content lives or dies on terminology. The ceiling is set by mic placement and how much room reverb you captured — then custom vocabulary closes the gap on drug names and acronyms. Numbers below come from real customer files.

95%
Lavalier or headset · single speaker CME

Solo presenter wearing a lav, recording straight to the device. With custom vocabulary on, drug names and acronyms hit reliably.

93%
Direct PA feed · conference talk

Recording pulled from the AV booth, not the room. Q&A from audience mics drops a few points; the speaker stays clean.

91%
Quiet-room research interview

Two speakers, condenser mic on the table, consented IRB-approved interview. Specialist terminology benefits from a vocab hint.

84%
Phone or back-of-room grand rounds

Phone propped on a podium or audience-recorded auditorium audio. Usable for search and review; expect a cleanup pass on drug names.

Common questions

8 things people ask about medical transcription.

01Can I upload a patient encounter or anything with PHI?+
No. We're not covered by a Business Associate Agreement and the platform isn't built for Protected Health Information. If audio identifies a patient — name, MRN, recognizable voice in a clinical encounter — use a HIPAA-covered service like Rev's BAA workflow, Abridge, or DeepScribe instead.
02What does count as a fit here, then?+
CME lectures and grand rounds you'd happily post publicly. Conference talks and panel recordings. Consented qualitative research interviews where participants signed off on transcription. Medical education content. Anything where there's no identifiable patient audio.
03I have IRB-approved research interviews. Are those OK?+
Yes, provided participants consented to transcription by a third-party AI service and the recordings don't contain other patients' PHI. Many IRB protocols specify this — check your approved consent form. We're a processor for non-PHI content; we don't sign BAAs.
04How well do you handle drug names and medical acronyms?+
Out of the box, the model knows common terminology — semaglutide, metformin, NSTEMI, COPD. For specialty work (oncology drug pipelines, gene symbols, trial acronyms) paste a vocabulary list on the job form and accuracy on those terms jumps considerably.
05How long do you keep the audio file?+
Audio is permanently deleted within 24 hours of the job completing. The transcript stays in your account until you delete it. This isn't HIPAA-grade retention policy — it's just our default for everyone.
06Can you transcribe a panel discussion with five speakers?+
Yes. Set speaker count to 4-8 on the job form and we tune diarization for shorter turns. Expect a 2-3 minute cleanup pass to rename the auto-labels (Speaker 1, 2, 3) to actual names. Works best when each panelist has a mic.
07Do you transcribe medical conferences in languages other than English?+
Yes — 99 languages with auto-detection, including Spanish, Mandarin, German, French, Portuguese, Japanese. Custom vocabulary works in any of them, so you can hand in Spanish drug brand names or German trial acronyms.
08What's the difference between you and Dragon Medical One or Abridge?+
Those are clinical workflow tools — ambient scribes or dictation engines that live inside the EHR and are built around patient encounters under a BAA. We're the opposite end: file in, transcript out, for the educational and research audio that doesn't belong in an EHR at all.

Drop your CME lecture or research interview. See what comes out.

30 free minutes every month. No card. Custom vocabulary, 99 languages, audio deleted in 24h. Non-PHI content only.

Start free