Every Raksa agent runs under a Mandate, passes the Human Gate, is watched by Rex & Kira, and leaves its tracks in the Evidence Ledger, all inside your Data Boundary. Not promises in a policy document. A part of the product you can open and check.
Every agent has a defined job.
A Mandate sets what an agent may work on and which of your policies it must follow. Asked to go beyond it, the agent declines and records the attempt.
Nothing is final without a person.
Everything an agent produces is a draft until someone on your team approves it. Every approval records who, what and when, in the app or in Slack alike.
Rex and Kira review the other agents.
Two agents whose only job is oversight. They check the others' work against your policies and flag anything that looks wrong. People stay above both.
Every agent action is on the record.
An exportable, tamper-evident record of everything the agents did. Ready for an audit, a regulator, or your own board asking what the AI did last quarter.
Your data stays yours.
Your data is never used to train any AI model. It is hosted in the EU and UK, kept separate from every other customer, and can stay entirely inside your own cloud with the Sovereign edition.
Anything the app can do, the API and agents can do, with the same guardrails attached. Wherever the work happens, the same approvals apply and the same record is kept. This is how it fits together:
Mandates, the Human Gate, the Watchdogs and the Evidence Ledger. Built into every layer, not bolted on.
Same product, same guardrails. The difference is where it runs.
For most teams
For regulated industries and strict data rules
Run a free assessment with no login, or book a walkthrough on your own scenario.