Security

Keep your data secure,
with permissions set your way.

Files stay where they are, permissions match the source systems and every access and action is logged. Important decisions always wait for you.

Your data stays put, and every step is visible

Files don't move, permissions follow the source systems and every access and action is logged.

Files stay put

Our servers keep only an encrypted index and read from the original location when needed.

Same permissions

If you can't see it in the source system, you can't see it through ANIMA.

De-identification

Optional. Names, phone numbers and similar fields become codes before reaching the model, then are restored in the reply.

Every step is logged

Logs who opened which file, when and what they did. Logs can be made tamper-proof.

First line of defense

Input and output,checked both ways

Blocks instructions hidden in documents and filters out anything that shouldn't appear in a reply.

Second line of defense

Tool accessset by role

The role decides which tools ANIMA can use. Tools that aren't allowed never show up in the list.

Third line of defense

Outside actionsneed your OK

Before ANIMA sends an email, writes to a shared sheet or takes another outside action, it lays out what it will do and waits for you to confirm.

Choose the setup your regulations call for

Mixed setups work too. How a private cloud, on-premises or mixed setup is put in place is assessed with you before rollout.

Cloud · ANIMA Cloud

Run by ANIWEAVE. No servers of your own, and the fastest way to get started.

Private cloud · VPC

ANIMA and the models run in your own cloud account. You control the network and storage.

On-premises

Chats, search and de-identification all happen in your own server room, managed by you.

FAQ

No. Your data is used only to serve your company, never to train models.

What's kept, and for how long, is set to your company's needs.

With the on-premises option, chats, search and de-identification all run in your environment. How a private cloud, on-premises or mixed setup is put in place is assessed with you before rollout.

We run regular red-team tests from an attacker's point of view, covering injected malicious instructions, sensitive data leaks and permission abuse. Anything we find becomes a new rule, and we test again whenever models or tools are updated. Your security team can review the test scope and related documents, and ask us to test specific items.

We go through five things with your team: where the models run (cloud, private cloud, on-premises or mixed), whether to turn on de-identification and which fields it covers, role permissions for each team and job, which actions need human approval (all outside actions by default) and who can view the audit logs (usually your security or audit team).

Put ANIMA to work for your team

Book a 30-minute call to find the work best handed to ANIMA.