Local AI on hardware you own, with applications shaped around your work. Start with clinical practice, or scope a workflow around the knowledge that makes your business yours.
One local foundation. A family of focused applications.
01 / The Sentinel family
Local intelligence is the common ground. Each application is shaped around the people, information and decisions it serves.
In use today
For clinical practice
The work around care, brought together. Records, scheduling, telehealth and billing, with local AI to help draft the notes you review.
Our most established configuration, shaped by a working clinical practice.
Read the Sentinel story on WealthBriefing (opens a new tab)Pilots starting
For family offices
Private document intelligence. It reads and answers from the family’s papers in plain English, citing the page behind every line.
Runs on one box in your office. No cloud AI.
See a sample answerDirections we’re exploring
Exploring controlled collaboration on sensitive data, with access and permitted outputs defined up front.
Exploring local research workflows for proprietary literature and scientific knowledge.
Exploring ways to work with study documents and research knowledge within defined data boundaries.
These are research directions, not products available to deploy. Scope and feasibility still need to be established.
Local models. Hardware you own. Data boundaries agreed for your deployment.
Explore everyday AI02 / A different starting point
Client files. Commercial strategy. The knowledge that makes your business yours. Using AI shouldn’t mean sending it all to someone else’s servers.
See where your data goesRun document retrieval and model inference locally. Decide what connects to the outside world, instead of making cloud AI the starting point.
Use your own compute for recurring workloads rather than paying a cloud provider per prompt. We scope hardware, setup and ongoing costs before you buy.
Choose models that suit your work and hardware. Keep local AI available without an internet connection once the system is set up.
03 / Everyday capabilities
The local AI tools behind the applications.
Find a starting point for your business.
Core capability
Ask questions across policies, research and project files. Work with draft answers grounded in your own documents, with references you can check.
For legal, finance, research and knowledge teams
Discuss this use caseWhat do we need before onboarding a new supplier?
A signed agreement, a completed security review and a named internal owner. Finance also needs the supplier’s tax and payment details.
Business workflows are scoped per deployment. AI output needs human review; model quality and capacity depend on the configuration.
04 / A clear boundary
The model and document retrieval run on your appliance. Your computers connect to it through your network. External services are a separate deployment decision.
A browser. Your own files.
Ready for your review.
Email, remote access, payments and off-site backups can involve external services. We define those boundaries with you. Local AI processing is not a promise that every feature is offline.
05 / From idea to installation
Your business comes first.
The configuration follows.
Choose your first use case, the people using it and the information that needs to stay local.
Scope hardware, models, document setup, external connections and any custom workflow.
Review the quote, availability, setup and support terms before making a purchase.
No. The deployed clinical-practice system is one configuration. The underlying local models, document retrieval and speech tools can support other business workflows. We scope and validate your use case rather than promise that every industry integration is ready-made.
Local inference and document retrieval can run offline after setup, with the required models and documents installed. Downloads, updates, remote access, email, payments and other connected features may need a network connection. Your deployment plan identifies those dependencies.
That depends on how much you use it, the models you need and your support requirements. Local inference avoids a cloud provider’s per-request model charge, but hardware, electricity, setup and maintenance still cost money. We compare those factors for your workload before recommending a configuration.
We select a configuration around your documents, response-quality needs and expected usage. Model availability, licensing, memory and speed all matter. Your quote specifies the hardware and model setup; the website illustration is not a fixed hardware specification.
No appliance can make an organisation compliant by itself. Local processing can support your privacy requirements, but compliance also depends on configuration, access controls, retention, contracts and how you use the system. We discuss these requirements when scoping your deployment.
We discuss your first use case, users, data boundaries and existing systems. The quote defines the hardware, model configuration, setup, any custom work and support terms. Pricing and availability are confirmed for your deployment before you commit.
Get Sentinel
Start with your work, not a hardware spec sheet. Request a tailored quote for your Sentinel deployment.