Security and trust

Written for the person who gets blamed if it goes wrong. Eight plain statements. If one stops being true we change this page the same day.

1. Approval before action

Restart, deploy, spend, delete, rotate: each becomes a card in your channel and in the pane that only a named owner or admin can press. The agent waits. Read-only work needs no approval and is still logged. Collaborators can ask; nothing they ask for runs without an owner's approval.

2. A vault the model never reads

Secrets are stored by name, encrypted at rest with AES-256-GCM under a platform key that lives only on the server. When a tool needs a token it fetches the value at the moment of use, server-side, and the fetch is logged as secret.used with the secret's name. The value never enters a prompt, a transcript, a memory or a browser. You see a name and a last-used time, never the value, not even your own.

3. One workspace per customer

Your agents run in their own isolated worker with their own limits and network rules. No customer shares a process with another. You can wipe yours without touching anyone else's.

4. Your key, your cap

Inference runs on an Anthropic key you create in your own Anthropic console, with a spend limit you set there and a monthly cap you set in the pane. Agents warn you at 80% of the cap. We never see your Anthropic bill and never carry it.

5. An audit you can export

Every job, approval and secret name used, with who pressed what and when. Download it as CSV from the pane at any time. It is yours.

6. Hosted where you choose

You pick an Olsie Cloud region when you sign up: Germany (Frankfurt) for EU data residency, or United Kingdom (London) for UK data residency. Your workers, their memory and your vault stay there. Servers are ours to run; the provider behind a region is an implementation detail we can change without moving your data out of the country. We are a data processor under UK and EU rules; a short data processing agreement is available before you sign.

7. What we cannot see

Your secret values, your Anthropic bill, and anything in channels you have not invited an agent into.

8. What is not built yet

Slack and Teams arrive after Discord. Email invites for collaborators are next. Workers start automatically within a minute of your Discord server and Anthropic key being in place; if either is missing the Team page says exactly what to add. We say so here rather than hint otherwise.

Reporting a problem

Email security@olsie.ai, or hello@kamobo.co.uk if that bounces while the address is being set up. We reply the same working day.