Ordeum
Skip to content
Back to Ordeum

Security Center

AI does the work. You stay in charge.

Data is masked, actions need approval, risks are blocked. Security doesn’t rely on the model. Ordeum checks every step.

Privacy Policy Terms of Service Security Center Acceptable Use Policy Refund & Cancellation

Two checkpoints

Before sending: we check for sensitive data. Keys and connection strings leave as placeholders. Public shares hide sensitive details automatically.

Before anything acts: key actions wait for you. Offline by default; each connection approved once, then revoked. Email and writes to outside systems run only after you approve. Destructive commands are stopped before they run.

Foundations

Encryption & key protection. Credentials are sealed with AES-256-GCM and bound to identity. Traffic is encrypted end to end; keys never reach the agent.

Data residency & governance. Deployed regionally in Singapore, with clear rules for storage, retention, deletion and cross-border transfer.

Zero-trust execution. gVisor sandboxes and least privilege keep access closed by default; every external action is authenticated and policy-checked.

International standards. Built to align with frameworks for information security, AI governance and service audits, from access control to activity audit.

Found a security issue? Email support@ordeum.ai with “Security” in the subject.

Safety & Permissions

Last updated: October 19, 2026

On this page
  1. Actions within your permissions
  2. Checks for supported text and actions
  3. Outcomes and review
  4. Privacy during security review

Ordeum’s permissions and safety controls help keep AI work within your authorized access and our Acceptable Use Policy.

Actions within your permissions

You choose your connections. Agent follows their permissions and your conversation’s approval settings. Some actions need separate confirmation; review the action details before approving. Broader permissions do not override platform restrictions or permit prohibited actions.

Checks for supported text and actions

Supported content checks block requests identified as prohibited from the first occurrence. Context and intent are considered when reviewing uncertain cases. Other checks can detect credentials, warn about personal information or require approval before an action. AI providers may apply additional safety rules.

Coverage depends on the feature and enabled protection; not every file, image or tool result is inspected. Your Security center shows settings and recorded outcomes. Missing records do not certify content or activity as safe.

Outcomes and review

Requests may receive warnings, restrictions or refusals. Repeated or serious misuse may restrict an account. Automated flags do not establish unlawful conduct; service failures do not imply misuse.

Request and track eligible reviews in your Security center. Report incorrect restrictions or harmful shared content to support@ordeum.ai. Include the relevant link or record ID, without credentials or unnecessary personal information.

Privacy during security review

Security review uses technical logs, risk signals and redacted information where sufficient. Staff access is limited to support you authorize, necessary security investigations or legal requirements under our Privacy Policy. Safety checks do not authorize routine staff browsing of your work.

For service availability and geographic restrictions, see our Terms of Service.