Security by design

Built so your data stays yours.

Security isn't a setting we turned on later. Every company gets its own environment, every person their own sandbox, and the AI never holds a single password.

How it fits together

Walls around your company. Walls around each person.

Eight promises we build on

  1. 01

    Your own environment

    Each company runs separately, with its own data, logins and settings. On a shared plan, only the servers and the operations tooling are shared, never your content.

  2. 02

    One sandbox per person

    Every person's agent runs in its own hardened sandbox, kept apart from their colleagues' agents.

  3. 03

    Keys never reach the AI

    Passwords and access tokens are held by a key keeper outside the agent and added only as a request leaves for your app.

  4. 04

    You decide what it can do

    For every app and every person: no access, read only, or read and write. Read only is the default, and the rules are enforced outside the AI.

  5. 05

    Your own sign-in

    Each company has its own login service, so your team's accounts live only in your environment.

  6. 06

    Your own AI account

    Each company uses its own account with the AI provider, or its own models on its own servers. Never a key shared between customers.

  7. 07

    Your connections, your apps

    Each company connects through its own app registrations, so access can be reviewed or revoked from your side at any time.

  8. 08

    Free to move

    Your environment is built to move as a whole: from our cloud to yours, or onto your own servers.

What the AI sees

To answer, the AI model sees the question and the data the agent fetched to answer it, with that person's access. It never sees passwords or keys.

If you run models on your own servers, nothing leaves your network at all.

Where your data lives

In our cloud, your environment runs in the European Union. In your cloud or on your servers, it runs wherever you put it.

See the deployment options →

Questions IT teams ask

Does the AI ever get our passwords?

No. Passwords and tokens stay with the key keeper, outside the agent. They are added only when a request leaves for your app, so the AI never sees them.

Can an employee see a colleague's data?

Only what their own access allows. Each agent works with that person's permissions in your apps and the access level you set for them in VectorLink.

Can the agent change things in our apps?

Only where you allow it. Every app starts as read only; you choose, per person, where the agent may also write.

What happens if we stop using VectorLink?

Your environment is yours. We can move it to your own cloud or servers, or shut it down; the app connections are yours to revoke from your side at any time.

Have a security question we didn't answer?

Bring your IT team to the demo. We'll walk through the setup and answer every question.

Cookie settings

Choose which kinds of cookies we may use. Optional tools are only loaded after you allow them.

Strictly necessary
Always on

Needed for the site to work and stay secure, and to remember your cookie choices. They don't track you.

Analytics
Google Analytics

Help us understand which pages are useful and how people find us. The numbers are combined; we don't use them to identify you.

Advertising
Google Ads, and other ad tools we list in the cookie policy

Measure whether our ads work and show our ads to people who visited the site.

Details of every cookie are in our cookie policy. We ask you again after 6 months.