The Jes family of local-first apps

JesType came first

JesType is an offline dictation app: you speak, it types, and the speech recognition runs on your machine rather than in someone else's data centre. It was built on a simple observation — the models had become good enough and small enough to run locally, so there was no longer a technical reason to send every sentence you dictate to a subscription service.

People who use JesType tend to use it for exactly the things they would not put in a cloud tool: patient notes, client matters, private drafts. That taught us what the product really is. Not "dictation with a privacy feature" — a tool you can use without asking permission from a procurement process.

JesRecap is the same principle, applied to meetings

Meeting recorders went the other way. The standard shape is a bot that joins your call, uploads the audio, transcribes it in the cloud, summarises it with a model you did not choose, and charges every user every month. It works well, and for a lot of conversations it is completely fine. For the other conversations, it is a non-starter — and those are the meetings where good notes matter most.

So JesRecap does the same job the other way around. It records your microphone and the audio your computer plays, as two separate tracks. It transcribes with a speech model on your disk. It writes everything into a searchable library on your machine and leaves it there. No bot joins the call, so it works with Zoom, Teams, Meet, Webex, Slack huddles, any app at all, or an in-person conversation through the microphone.

It deliberately stops short of one thing: there is no cloud AI in it, and there never will be. When you want a summary or a list of action items you have two routes, and neither involves a model vendor. It can copy a ready-made prompt with your transcript into the clipboard for whichever assistant you already trust, or — if you turn on the optional AI notes setting — write the note itself using a model that runs on your own machine. See meeting actions. We are not going to route your transcripts through somebody's servers and describe it as a feature.

How we work

Small, independent, and slow in the way that small teams are slow: features ship when they are right rather than when a roadmap slide says so. Support is an email address, read by a person who can change the code. If you write with a real problem you will get a real reply, not a knowledge-base article.

We do not publish user counts, star ratings, or testimonials. Not out of modesty — we just do not think numbers you cannot verify should influence your decision. Download the preview and see whether the recording quality and the transcript are good enough for your work. That is the only review that matters.

Get in touch

Support, questions, security reports, licensing for a team, or a straight answer to a security questionnaire: support@jesrecap.com. For fleet deployments, start with the enterprise page and the deployment guide.