Your meetings, written down and never uploaded
JesRecap is a meeting recorder for Mac and Windows. It records your microphone and the other side of the call as two separate tracks, transcribes them on your own processor, and leaves a searchable transcript on your disk. Nothing joins the meeting, and nothing is sent anywhere.
The complete app for 48 hours. No card, no account, nothing gated.
Two files, written to your disk as you talk. Neither one leaves it.
Both land in a folder you can open yourself:
~/Library/Application Support/jesrecap/meetings/
Recording a call in Zoom, Teams, Meet or Webex. Or the person across the table.
A transcript you can search, quote and hand over
Stop the recording and the transcript is there: timestamped, attributed, and yours to edit. Nobody has to sit through the call twice to find the one thing that was agreed.
- You04:12
- Before we sign anything, I want to be clear on what happens if the pilot slips past November.
- Speaker 204:21
- Then the second invoice moves with it. We only bill on the milestone, not on the date.
- Speaker 304:38
- I can put that in writing this afternoon so nobody has to remember this conversation.
- You04:44
- Please do. And add the two names we agreed on for the review board.
- Speaker 205:02
- One thing I want on the record: the security review has to clear before we migrate anything, not after.
- You05:11
- Agreed. Nothing moves until that comes back signed off.
Billing is tied to the pilot milestone rather than the November date, so a slip moves the second invoice with it. Written confirmation was promised for the same afternoon, along with two names for the review board.
- Confirm milestone billing in writing, Speaker 3, this afternoon
- Name two reviewers for the board, agreed at 04:44
- Re-check the November pilot date before signature
notes.mdtranscript.txtmeeting.jsoncaptions.srtcaptions.vttmeeting.wav
Every meeting exports to all six. Formats that will still open in ten years, on a disk you control.
Four screens, and you have seen all of it
Record, transcribe, label the speakers, turn it into notes. The tour moves on its own; hover to pause it, or click a tab and take over.
Your voice and their voice are never on the same file
Most recorders capture one mixed stream and then try to pull the speakers apart afterwards. JesRecap keeps your microphone and the meeting audio separate from the first second, and only mixes a playback copy once the meeting is over. That one decision is why attribution holds up, why you can re-run a transcript later, and why you can hand over the other side of a call without handing over your own.
Attribution that cannot drift
You were never on the other track, so your words cannot be credited to someone else. The diarization model only has to separate the room, not separate you from the room.
Transcribe it again, better
The original audio stays on disk, so a meeting from March can be re-run through a larger speech model in December. Nothing about a transcript is final.
Export one side or both
Send the client the mixed audio, keep your own track back, or export nothing but the text. The WAV files are ordinary files in an ordinary folder.
Nothing to connect, nothing to configure
JesRecap records the audio your computer is already playing and the microphone you already use. There is no plugin, no calendar permission, and no participant to admit, which is exactly why it works in every meeting app and across a table.
Press record
Before the call or halfway through it. macOS hands over system audio through Core Audio process taps; Windows uses WASAPI loopback. Both are built in. There is no virtual audio driver to install.
Have the meeting
Two tracks are written to disk continuously while you talk. If the laptop dies mid-sentence, everything up to that second is still on the disk and still transcribable.
Read it back
Stop, and the transcript appears, written on your own processor. Speaker labels follow in the background. Then take the summary, copy a prompt into your own AI, or export and get on with your day.
€34.95 once, and then nothing
There are no servers transcribing your audio, so there is no monthly bill to pass on to you. Recording a thousand hours costs exactly what recording one hour costs, because the work happens on hardware you already bought.
Same job, opposite architecture
Otter, Fireflies, Granola and Read AI are capable products built on the assumption that your conversations can be processed on someone else's servers. Everything below follows from starting at the other end.
| JesRecap | Cloud meeting assistants | |
|---|---|---|
| Where audio is processed | On your device | Vendor servers |
| Bot joins the meeting | Never | Usually |
| Account required | No | Yes |
| Works with no connection | Yes | No |
| Original audio kept | Both tracks, on your disk | In the vendor workspace |
| Pricing | From €34.95 once | Monthly subscription |
The things worth knowing first
It is a preview, and we say so
Recording, transcription, labelling and export all work. Interfaces and defaults still move between builds. Every preview build is free to license holders. See the release notes.
Mac and Windows only
macOS 14.2+ on Apple Silicon, Windows 10 and 11 on x64. No Linux build: the microphone side is portable, the system-audio side is not.
Recording people has rules
Consent requirements vary by country and by profession. JesRecap hides nothing from anyone, but it also announces nothing: telling the room is on you, and we would tell them every time.
Local does not mean never online
Models download once, licenses activate and re-validate, and the app checks for updates. A paid license keeps recording for up to 30 days between successful checks. Your audio is never part of any of it.
Summaries are yours to choose
Turn on AI notes on this device and a local model writes them here. Leave it off and one click copies a prompt with the transcript, for whichever AI you already trust. See meeting actions.
No team features yet
There is no shared workspace, no calendar sync and no integrations. If you need a searchable library your colleagues can read, a cloud notetaker is the honest answer.
Your meetings are not anyone's training data
Recording, transcription and speaker labelling all happen on your machine. The network is used for model downloads, license activation and validation, update checks, and any calendar connection you choose to set up. Whether meeting content ever goes anywhere else is a decision you make, one export at a time.
Work where the conversation cannot be handed to a third party
Local processing is not a preference in these rooms; it is usually the only version of the tool that gets approved.
Clinicians and therapists
Session notes without a business associate agreement, because there is no third party to sign one with.
Healthcare and psychologyLawyers and advisers
Privileged conversations that never leave the office network, with both audio tracks kept as the record.
LegalFinance and public sector
Client and citizen meetings recorded inside your own perimeter, under your own retention rules.
Financial services and governmentFrequently asked questions
Does JesRecap work with Zoom, Teams, and Google Meet?
Yes, and with every other app that plays audio through your computer, including Webex and Slack huddles. JesRecap captures the system audio directly, so there is nothing to integrate and no plugin to install. In-person conversations work through the microphone alone.
Does a bot join the call?
No. Nothing is added to the meeting and no extra participant appears in the list. JesRecap records the audio your computer is already playing, so the call looks completely ordinary to everyone else in it.
Does anything ever leave my device?
Your recordings and transcripts do not. JesRecap uses the network for four things only: downloading speech models once, activating and periodically validating your license, checking for updates, and any optional calendar connection you set up. Content you export, back up, or paste into another service follows that destination's privacy practices.
How accurate is the transcription?
JesRecap runs speech models on your own machine, and you choose which one. Larger models are more accurate and a little slower. Because both original audio tracks are kept, you can retranscribe any meeting later with a different model without recording it again.
Can it tell who said what?
Yes. Your own voice is separated by design, because your microphone is recorded as its own track. The other voices are separated after the recording by a local diarization model, a one-time 54 MB download. You can rename any speaker, and JesRecap will not overwrite a name you typed. More on speaker labels.
Where do the summaries come from?
Two ways, both under your control. Enable AI notes on this device and a local model writes the summary on your own machine (a one-time download of about 530 MB). Leave it off and the same buttons copy a ready-made prompt with the transcript included, for you to paste into whichever AI you already use. There is no cloud AI in either case.
Is there really no subscription?
Really. A one-device license is €34.95; 2, 4, and 8-device licenses are €49.95, €74.95, and €99.95, each a single payment. There are no servers processing your audio, so there is no monthly bill to pass on and no per-minute meter anywhere in the product.
Can I try it before paying?
Yes. The complete app runs for 48 hours after you verify your email address with a six-digit code in the activation screen. No card and no account are required, and nothing is gated during the trial. How the trial works.
Do other people in the meeting see that I'm recording?
JesRecap adds no bot and shows nothing inside the meeting app. Recording conversations may require consent where you live and work. We recommend simply telling people, every time. It's also just good manners.
Which operating systems does it support?
macOS 14.2 or later on Apple Silicon, and Windows 10 or 11 on x64. There is no Linux build today: microphone capture is cross-platform, but the system-audio side is written against the macOS and Windows audio APIs.
Record your next meeting and see what you have been missing
The complete app for 48 hours, on Mac or Windows. No card, no account. After that it is €34.95 once, and never again.