JesRecap for organisations
A meeting recorder that adds no server, no tenant and no data-processing relationship. Everything happens on the endpoint you already manage — so the review is about one desktop app, not an entire cloud platform.
Volume licensing
Licenses for a team, a department, or everyone. Pricing is agreed directly rather than published, and there is no per-minute or per-meeting component to forecast.
Nothing to host
There is no JesRecap server, no container to run, no database to back up. The app is the deployment.
Standard installers
A macOS .dmg and a Windows .msi that behave like any other package in your existing deployment tooling.
No accounts to provision
No SSO integration, no user directory sync, no seats to reconcile — because the app has no login at all.
The architecture, for your security review
The short version: meeting content never travels. The only outbound traffic is a one-time model download and license activation.
| Component | Where it runs | Leaves the device? |
|---|---|---|
| Microphone capture | On the endpoint, cross-platform audio input | No |
| Meeting-audio capture | On the endpoint — Core Audio process taps on macOS, WASAPI loopback on Windows | No |
| Speech-to-text | On the endpoint, Whisper-family model | No |
| Speaker labelling | On the endpoint, local diarization model | No |
| Storage | Local WAV files and a local SQLite database | No |
| Summaries and action items | On the endpoint with the optional AI-notes model, or a prompt is copied to the user's clipboard | Not by JesRecap |
| Model downloads | models.jesrecap.com | Outbound HTTPS, once per model — speech models, plus the AI-notes model if enabled |
| License activation | JesRecap license service | Outbound HTTPS, license key and device registration |
| Telemetry, analytics, crash uploads | None | No |
On summaries: JesRecap never transmits a transcript, and there is no cloud AI in it. Its meeting actions have two modes. By default they copy a ready-made prompt, with the transcript included, to the clipboard; if a user then pastes that into an external assistant, that transfer is their action inside whatever tool your organisation has already approved — worth naming explicitly in your policy. The alternative is the optional AI notes on this device setting, off by default, which generates the summary with a local model so the transcript stays on the endpoint. Both are worth a line in the policy.
What this removes from the review
Most of the work in approving a meeting-notes tool is about the vendor's cloud. There isn't one here.
No data-processing footprint for content
We never receive your recordings or transcripts, so there is no retention window to negotiate, no sub-processor list for meeting content, and no deletion request to file.
No bot in the meeting
Nothing joins the call and no meeting platform needs to be configured or trusted. JesRecap records the audio the endpoint already plays, which is why it works with Zoom, Teams, Meet, Webex, Slack huddles or anything else without an integration.
No residency question
Recordings live wherever the laptop is. There is no region to choose, because there is no storage on our side to place. See data sovereignty.
Designed to support your compliance process
On-device processing fits GDPR- and HIPAA-style requirements where cloud transcription is difficult to approve. We describe the architecture; we do not claim a certification. The GDPR and HIPAA guides go into detail.
Predictable cost
A one-time license per user rather than a monthly per-seat bill, with no usage meter. Budget once and forget about the renewal cycle.
Egress you can enumerate
Two hostnames, both HTTPS, both optional after first use: the model host and the license service. Easy to allow, easy to audit, easy to explain to a network team.
Deployment
Three steps, and none of them involve infrastructure.
Evaluate
Install the preview on a couple of managed machines from the downloads page and put it through your usual endpoint and network review.
License
Agree a volume arrangement with us by email and receive keys for your users. No portal to onboard, no directory to connect.
Roll out
Push the .dmg or .msi through the tooling you already use. Allow the model host and license service through the proxy, and users are done — the first recording prompts for audio permission on macOS and microphone access on Windows.
IT and security questions
Which outbound connections does JesRecap make?
Two. It fetches models from models.jesrecap.com — the speech models, plus the roughly 1.3 GB AI-notes model if a user turns that optional feature on — once each, then reuses the local copy. And it contacts our license service to activate a key and register the device. No meeting audio, transcript or metadata is included in either. There is no telemetry or analytics channel.
Can it work on an isolated machine?
Once a model has been downloaded and the license activated, recording, transcription and speaker labelling need no network access at all. Those two initial steps do require connectivity.
Where is meeting data stored on the endpoint?
In JesRecap's own application-support directory: WAV files for the microphone track, the system-audio track and the mix, plus a SQLite database holding metadata and transcripts. It is covered by your existing disk-encryption and backup policy like any other local file, and deleting it deletes the only copy.
Does anything get installed at driver or kernel level?
No. macOS capture uses the Core Audio process-tap interface the OS provides, gated by the standard Audio Capture permission the user grants once. Windows capture uses WASAPI loopback, which is built into Windows. No virtual audio device is added.
Is there an admin console or central policy control?
Not today. JesRecap is a per-user desktop app with no server component, so there is no fleet dashboard, no central transcript store and no remote configuration. That is a deliberate consequence of the architecture — worth knowing before you plan a rollout.
Do you offer SSO or directory integration?
No, because the app has no accounts to sign into. Entitlement is a license key, not an identity.
How do we handle staff turnover?
A license covers a fixed number of devices — 1, 2, 4, or 8, depending on the tier — and device assignment can be managed from device management. For volume arrangements, get in touch and we will work out reassignment with you.
Can we get documentation for our vendor-assessment questionnaire?
Yes. Start with the security page and privacy, then email support@jesrecap.com with whatever your process still needs answered.
What does the preview status mean for a deployment plan?
JesRecap ships on a preview channel rather than a settled 1.0. Recording, transcription, labelling and export all work, but interfaces and defaults still move between builds. Treat it as an evaluation you can already do real work in, and track the releases page.
Is Linux supported?
Not today. Microphone capture is cross-platform, but system-audio capture is implemented against macOS and Windows audio APIs, so those are the two supported platforms.
Fewer questions to answer
No server, no tenant, no meeting data in transit. Tell us how many people need it and we will sort out licensing.