QuickChart ProximitySafety & Privacy Architecture
A hybrid AI infrastructure that processes on your own device where possible, then on QuickChart's self-managed GPU cluster inside our own servers, and uses subprocessors only for certain tasks.
Built by Canadian specialists - deployed across thousands of clinics and trusted in millions of visits.
- Mostly stays on device
- Some to our GPU cluster
- Certain tasks to subprocessors
Safety Focused AI Infrastructure
Each AI task is handled as close to you as it can be done well. Data only moves outward when the task genuinely requires it.
On-device processing
On supported hardware, such as a MacBook Pro with Apple M4, eligible AI tasks are processed locally on the clinician's own device. When a task can be handled on-device, that data doesn't need to leave the room to get the job done.
- Runs on supported Apple silicon (MacBook Pro M4)
- Eligible tasks never leave the device to be processed
- The closest possible point of processing
Most AI tools send everything out. We start close to home.
Many AI documentation products rely entirely on third-party processing. Proximity flips the default: local where possible, our own infrastructure where possible, and subprocessors only for the tasks that need them.
Other Scribes
A common approach across the market
- Your deviceCaptures the data
- Third-party AI providerEvery task is sent out
- Back to the appResults return
QuickChart Proximity
Hybrid, closest compute first
- Your deviceProcesses eligible tasks locally
- QuickChart GPU clusterOur own infrastructure
- SubprocessorCertain tasks only
The same protections, whichever layer does the work
Encrypted at every layer
Data is encrypted in transit and at rest, wherever a task is processed.
Never used to train AI
We don't train AI models on your clinical content, and approved processors are contractually bound by no-training commitments.
Identifiers reduced before leaving
Where technically feasible, identifier-reduction is applied to clinical text before it's sent to an approved external AI processor.
Stored in your region
Wherever a task runs, your records are returned to and stored in your selected region.
Common questions
What does “where possible” mean?
Not every task can run on a laptop or in our cluster today. Proximity routes each task to the closest layer that can handle it well, and only moves outward when a task requires it.
Do I need a MacBook Pro M4 to use QuickChart?
No. On-device processing is used where supported, such as on a MacBook Pro with Apple M4. On other devices, tasks are handled by QuickChart's own GPU cluster and, for certain tasks, subprocessors.
Which tasks use third parties?
Only certain specialized tasks. Approved processors are bound by no-training commitments. See our Compliance & Security page and Privacy Policy for full details on processors and processing modes.
AI that stays as close to you as possible
Try QuickChart and see the Proximity architecture at work in your practice.