RU

Gabbi: session notes in a minute — and they never leave your laptop.


A local AI that drafts your psychotherapy notes on your own machine. No cloud, no third parties: client confidentiality by architecture, not by policy.

Get early access

Sound familiar?

The after-session hour

Ten to fifteen minutes per client reconstructing the session and writing it up. Across a full caseload that's an hour a day you don't bill.

A patchwork you can't vouch for

Phone recorder, a transcription service, then ChatGPT, then edits — client material scattered across vendors you've never audited.

Confidentiality you can attest to

You can quote a cloud tool's encryption, but you can't prove to a client that their voice never left your control. With Gabbi, you can.

How it works

Dictate or paste a transcript

Right after the session, dictate the key points or drop in a transcript.

The on-device AI drafts the note

In about a minute it structures the material — presenting issues, progress, hypotheses, plan, next session — without a single network call.

Review, sign, export

Adjust the wording, sign off, and export to PDF or DOCX in one click.

Why on-device matters

Your laptop is the only server. No cloud, no third-party processors, no network egress — so there's no data-processing agreement to sign, no sub-processor list to track, and nothing to breach remotely. Notes are encrypted on disk. It works on a plane, on a train, in a room with no Wi-Fi. This isn't a setting; it's the architecture.

Early access

€79 / mo

Founding members get the full version before launch, a direct line into how it develops, and a perpetual-licence option.

Join the founding cohort

We're building Gabbi with practising clinicians who take confidentiality seriously. Leave your email to join the founding cohort — and if you're up for it, we'd value 20 minutes to take apart one anonymised note together.

No spam — just your invitation, and a note teardown if you want one.
This site uses cookie-free analytics and a secure waitlist form. The Gabbi desktop app itself runs fully offline, with zero network egress.