Meeting notes for work that can’t go to the cloud

Why confidential work needs on-device transcription

Most meeting assistants work the same way: a bot joins your call, or an agent streams your audio to a server, and the transcript is produced somewhere else. For a lot of teams that is fine. For anyone handling patient details, privileged conversations, client finances, citizen data or student records, it means a new processor in the chain and a new set of questions — where is the audio stored, for how long, who can read it, which country is it in, and what happens if the vendor changes its terms.

JesRecap removes the question instead of answering it. Capture, storage and transcription all happen locally: macOS uses Core Audio process taps, Windows uses WASAPI loopback, and a Whisper-family speech model runs on your own CPU or GPU. The model is downloaded once from models.jesrecap.com and reused after that. Recordings live in a folder on your disk as WAV files with metadata in a local SQLite database.

Two smaller details matter just as much in sensitive settings. First, nothing joins the meeting: JesRecap records the audio your computer is already playing, so participants see no extra attendee and no recording bot in the list. Second, there is no cloud AI anywhere in it. When you want a summary you press a Meeting action, which copies a ready-made prompt with the transcript to your clipboard so you can paste it into whichever AI your organisation has already approved — or into nothing at all. If you would rather not involve an assistant at all, the optional AI notes on this device setting has those same buttons write the note with a model that runs on your own machine.

What every industry below gets

• Two separate tracks — your microphone and the system audio — plus a mixed playback track created after the meeting.
• Local transcription, with an optional live transcript while you are still talking.
• Speaker labels added after recording by a local diarization model (a one-time 54 MB download).
• Exports for the file: Markdown notes, plain text, a JSON archive, SRT/VTT captions, and the WAV audio (mixed, microphone-only or system-audio-only).
• A searchable meeting library, and the ability to retranscribe any old meeting later because both original tracks are kept.
• From €34.95 once, no subscription. Licence tiers cover 1, 2, 4, or 8 devices for the same person.