Every format a recorder produces
M4A from Voice Memos and most phones, MP3, WAV, AIFF, OGG and Opus. Video files work too — MP4 and MOV go in directly, with no step to extract the audio first.
Voice recording transcription
Transcribe a voice recording without uploading it.
Drop in a voice memo, a lecture, an interview, a note you left yourself in the car. Kekoso turns the recording into text on the Mac it is sitting on — with timestamps, in up to 99 languages, and with nothing sent anywhere to do it.
macOS 14.4 or later · Apple silicon · $29 once, not per month · 7-day trial
How it works
From a voice recording to a transcript in three steps.
No account to create, no file to upload, no queue to wait in. The recording is on your Mac, and so is the model that reads it.
- 01
Get the recording onto the Mac
AirDrop it from your phone, drag it out of Voice Memos into Finder, or point Kekoso at a folder your recorder already saves into. Any file macOS can play is a file Kekoso can transcribe, so to transcribe audio recording to text nothing has to be converted first.
- 02
Drop it in
Drag the file onto the Transcribe screen, or right-click it in Finder and choose Open With → Kekoso. Recognition starts on your own machine, with the model already loaded, and nothing is queued anywhere else.
- 03
Take the text wherever it goes next
Copy the transcript, or export it as TXT, SRT or VTT with timestamps. Hand it to your notes app, your editor, or your AI agent through the local MCP server — the recording itself stays where you put it.
What it takes
Any recording, any length, any of 99 languages.
Voice recording transcription usually arrives as a service with a free tier that runs out on the second file. Here the limit is your disk.
Hours long, no meter running
A two-hour lecture or a three-hour interview transcribes the same way a thirty-second note does. Length is limited by your disk, not by a plan tier, because nothing is billed per minute.
Timestamps you can use
Every transcript carries timecodes, so a quote leads straight back to the moment it was said. SRT and VTT export makes a recording of a talk into subtitles for the video of it.
A folder that transcribes itself
Point Kekoso at the folder your recorder syncs into and every new file is queued as it arrives. Twenty interviews from a field trip become twenty documents without twenty drags.
The names in your recording, spelled right
Add a client, a drug, a product or a colleague to the vocabulary once, and every transcript after that writes it your way. Fixes for words the model already hears, replacements for the ones it does not.
Up to 99 languages, chosen per model
Parakeet for fast English and 24 other European languages, Whisper for the long tail, SenseVoice for Chinese, Cantonese, Japanese and Korean. Each model shows its measured accuracy, so you pick the trade-off rather than trust an adjective.
Where your voice recording goes
Nowhere. That is the point.
A voice recording is usually the least guarded thing a person owns and the most revealing — a patient, a source, a client, a family argument, your own half-formed thoughts. Every online transcription service starts by asking you to hand it over.
Kekoso runs the speech model on the Mac. The transcription pipeline makes no network calls, the file is never copied, and the transcript lands in a plain local file you can open, back up or delete. Turn off Wi-Fi and transcribe something: it works exactly the same, because there was never a server in the loop.
How to verify that yourselfWhat it does not do
The limits, up front.
One microphone, one track
Kekoso does not label speakers in a recording made on a single microphone. A two-person interview comes back as one continuous transcript with timestamps. Where you record a call on the Mac itself, your side and theirs are captured as separate tracks — but a phone on a table is one track, and it stays that way.
The microphone decides more than the model
A phone on the far side of a meeting table produces audio no model recovers cleanly, and swapping to a bigger model will not fix it. Close, quiet and one voice at a time is what makes a transcript good; the software only works with what was captured.
Apple silicon only
macOS 14.4 or later · Apple silicon. Recognition leans on the Neural Engine, which Intel Macs do not have. Every MacBook Air and MacBook Pro sold since late 2020 qualifies, along with the Mac mini, Mac Studio, iMac and Mac Pro.
Questions
Before you drop the first file in.
- Can it transcribe a voice memo from my iPhone?
- Yes. Voice Memos syncs to the Mac over iCloud, or you can AirDrop the recording across; either way it arrives as an M4A file, and Kekoso takes that directly — the route to a transcript of voice recording made anywhere on the phone is the same one. The same goes for any other recorder app on the phone, as long as it can share the file.
- Does Voice Memos on the Mac not already do this?
- On macOS 15 or later with Apple silicon, Voice Memos can show a transcript of a voice recording made inside the app, and for reading back one note that is enough. It transcribes only recordings that live in Voice Memos, gives no timestamps you can use elsewhere, exports no subtitles and has no vocabulary to teach it names. A file in Downloads, a video, or twenty files at once are outside what it does.
- Does the recording get uploaded anywhere?
- No. The speech models run on your Mac, and the transcription pipeline makes no network calls at all. Turn off Wi-Fi and transcribe a file — it works the same, because there was never a server involved. The file stays wherever you put it; Kekoso does not copy it.
- How long can a recording be?
- As long as your disk allows. There is no per-minute charge and no monthly cap, so a three-hour recording costs the same as a thirty-second one: nothing beyond the licence. Long files take longer, and Kekoso keeps going while you do something else.
- Will it tell me who said what?
- Not for a recording made on one microphone. Kekoso does not do speaker diarization, so a conversation captured by a phone on the table comes back as one transcript with timestamps. Calls recorded on the Mac itself are different: your microphone and the other side are captured as two tracks and labelled as such.
- What formats can I export?
- Plain text, SRT and VTT, all with timestamps. A transcript of a voice recording leaves the app as a file you own, not a page on someone else’s site. Copy the transcript straight from the app for a quick paste, or export a file for subtitles, an editor or an archive.
- What does it cost?
- $29 once, for up to 3 of your own Macs, with free updates through the 1.x line. A 7-day trial unlocks everything, with no card and no account. Voice recording transcription is one part of the app — dictation, call recording and the MCP server are the rest, and all of it is in the same purchase.
- What languages does it handle?
- Up to 99, depending on the model you choose. Parakeet covers English and 24 other European languages quickly; Whisper covers the long tail; SenseVoice is there for Chinese, Cantonese, Japanese and Korean. The recognition language is set per model, or left on Auto.
Wondering whether the free route is enough? What Voice Memos transcription needs and where it stops is written out in transcribing a voice memo on a Mac. Weighing this against an online service or running Whisper yourself? The comparison is written out, costs and all, in how to transcribe audio and video to text on a Mac. If the file you have is a video rather than a recording, that has its own page: video to text.
Pricing
Buy it once. It's a tool, not a tenancy.
Four jobs that usually arrive as four subscriptions, running on hardware you already paid for.
$29once
7-day free trial. No card to start.
- Dictation, files and links, call recording and the MCP server — all of it
- No per-minute billing, no monthly cap, no usage dashboard
- Use it on up to 3 of your own Macs
- Free updates through the whole 1.x line
- Works offline indefinitely; no license check to phone home weekly
- 30-day refund, no argument, no questionnaire
macOS 14.4 or later · Apple silicon · payment handled by Lemon Squeezy
Try it on your own words.
7 days, every feature unlocked, no card and no account. If it does not earn its place in your week, delete it and nothing follows you.
macOS 14.4 or later · Apple silicon