Trust
Controls you can use, records you can pull
Human gates, pause, spend you can check, and your own model keys — built into the product.
Human on gated work
Work that needs sign-off waits until someone approves or denies. You can see which agents are waiting on a person. Decisions land on the activity trail.
Pause and readiness
Pause agents that need follow-up. Connected tools are visible, and an agent is not treated as ready for gated work until required connections are in place.
Spend you can check
Cost and tokens by agent. Runs without usable spend evidence show up as flagged gaps; nothing quietly looks free.
Your data and keys
Bring your own model API keys. We do not train models on your customer content. SOC 2 Type II audit underway.
The record behind every decision
Activity and evidence answer “what ran and who approved it” without asking an engineer to reconstruct it.
Vulnerability disclosure
Report a security issue to security@deixic.com. We acknowledge new reports within three business days.
While you test, do not access another customer’s data and do not degrade the service for other users.
We do not run a bug bounty program today.
If you report a vulnerability in good faith and stay within this policy, we will not pursue legal action against you for that research.
Documents
Subprocessors · Status · security.txt · Vulnerability disclosure · Privacy · Terms · SOC 2 Type II: audit underway
Subprocessors: the current list. Data processing addendum: DPA. Live app reachability: status. Longer essays on identity and containment: blog, thesis. This page is not legal advice.