For privacy and data protection
Personal data stays in the task that needs it.
Agents read whatever helps them finish the job. Kyvvu makes sure that personal data an agent reads doesn't travel on to places it shouldn't.
The problem
Purpose limitation for autonomous agents.
- 01
Agents read broadly
To be useful, agents get access to mailboxes, CRMs and file shares. What they read in one step can end up anywhere in the next.
purpose limitation
- 02
Hidden processors
An agent that calls an outside API with personal data has just created a transfer nobody assessed.
processors/transfers
- 03
Rights handled by machines
Erasure and access requests handled by agents still have to be legitimate, scoped and recorded.
data subject rights
In practice
Privacy rules, applied step by step.
Personal data to an outside tool
A support agent opens attachments with personal data and tries to send them to an outside transcription service.
Illustrative scenarios. Each one is decided by a single policy from the open policy library, as the engine evaluates it. References to laws and regulations are our reading of them, to show how a policy maps to an obligation. Not legal advice.
How Kyvvu helps
Data protection that follows the data.
- 01
Follows your classification
When an agent reads data classified as personal, the task remembers it. Outbound sends are blocked for the rest of that task.
data.classification: pii
- 02
Data subject rights
The GDPR pack requires approval before erasure and before automated decisions on personal data, and limits how much a task may read.
GDPR pack
- 03
Runs inside your environment
The engine sits in the agent's own process: so there's no proxy or extra network hop, and a check takes under a millisecond. Prompts and data stay with you.
in-process/no proxy
- 04
Every agent on the register
Each agent registers with its purpose, owner, risk classification, the tools it may use and its environment. Registering is free.
agent registry
Questions
Common questions.
01Is Kyvvu itself a processor of our personal data?
The engine decides inside your agent's process, so prompts and data stay there. Traces of steps go to the platform unless you keep them on your own machines, which one setting does. See the trust center for what runs where.
02Does it help with a DPIA?
Each agent's registration records its purpose, owner, the data it uses and the systems it touches, which is useful input. The assessment itself stays yours.
03Can it limit how much an agent reads?
Yes. The GDPR pack caps resource access per task, so an agent can't sweep up far more than it needs.
Get started
Every agent has to ASK before it acts.
Start with one agent
See it on your own agents
A 30-minute demo with someone who builds Kyvvu: your agent, your rules, and the policy that decides each step.
Read up
What runs where, and what leaves
The trust center explains what the engine does in your process, what reaches the platform, and our DPA.
Want to see other teams and sectors? All use cases.