Security & trust
Built so your security review can actually say yes.
There is one place inside your own systems where the rules are set and enforced. The data stays yours, the keys stay yours, and every action that matters leaves a record your auditors can replay.
LAST UPDATED · 13 JUNE 2026
It runs inside your own walls
Agents run where your data already lives: your own cloud account, your own datacenter, a specific country or region you name, or a device in the field. Nothing sensitive is copied to a Proqtor account to make the product work. The agents run inside your walls, and those walls are yours. We build each deployment to fit your existing infrastructure, privacy policies, data-sharing agreements, and the regulations you operate under.
- Your private cloud (VPC)
- Runs in your account. No data ever leaves to us. Your keys, in your key vault.
- In your building
- Runs in your datacenter. Can run fully offline.
- A specific country or region
- Locked to a US or EU region. Your data is guaranteed to stay there.
- On a device in the field
- Runs on the device itself. Works offline for cameras, robots, and field work.
One place to set the rules, and enforce them
Every request to an AI model passes through a single checkpoint inside your own systems. That checkpoint is where your rules live, which means there is one place to review and one place to enforce, not a tangle of settings scattered across every app.
- Clear control over who can access what, on every action and tool.
- A spending cap and a speed limit per workflow, with hard cutoffs.
- Personal details are masked before anything ever reaches the AI model.
- Every change to the instructions is recorded and can be rolled back.
- Works with any AI model, and can fall back to a private one you run yourself.
The agent only sees what the person is already allowed to see
When an agent looks things up in your own documents, it only ever pulls what that person or workflow already has permission to open. The same access rules apply file by file, so the agent can never become a back door that leaks what someone could not otherwise see.
You keep the proof it works
The full record of every step, every run, and every quality check is written into your own environment, kept as long as you choose. It is what you use to fix problems, to check quality, and to show an auditor exactly what an agent did, in what order, and which documents it used.
- Encryption
- You hold your own keys. Data is encrypted at rest and in transit, in your own key vault.
- Sign-on
- People sign in through your own company login. No separate password to manage.
- Who gets access
- We get the least access needed, only when you approve it, for a set time, and it is all logged.
- Audit trail
- Every action that matters, and every approval, is recorded and can be replayed.
Governed by design, not by promise
The EU AI Act treats workplace monitoring as high-risk, so Proqtor is built to stay on the right side of it. We watch the work, not the worker: we never score individuals and run no emotion recognition. Each agent follows a step-by-step path, earning more responsibility only as its own record proves it is safe, from helper to working on its own. A person stays accountable for the actions that matter.