Is OpenTrust ready to launch?
This page shows how OpenTrust keeps AI agents safe from bad tools. Try the tool simulator, review the launch checklist, and test the live site.
Try it yourself
Pick a tool. OpenTrust will decide if your AI agent is allowed to use it.
Tool info
Hello Weather
- Trust level
- Checked by the community
- Creator
- OpenTrust Demo
- Verified
- GitHub verified
- Price
- Free
- Network
- None
- Version
- 1.0.0
Decision
Allowed
This tool passed all checks. Your agent can use it safely.
What can this tool do?
Red means the tool has that permission. Green means it does not.
What happens to your data?
Launch checklist
These are the things OpenTrust checks before it goes live.
Tools have tamper-proof seals
If anyone changes a tool after it's signed, we know right away.
Blocked tools stay blocked
Once we remove a tool, AI agents can't use it — even when offline.
Dangerous actions are blocked by default
Wallet access, terminal use, and private data are off unless you say so.
Spending limits are enforced
Your agent won't spend more than your cap. Big payments need approval.
Payment codes expire quickly
Each payment code is single-use and expires fast so it can't be stolen.
Admin actions are logged
Every change to the registry is recorded with who did it and when.
The server is locked down
Security headers, rate limiting, and non-root containers are all on.
Real keys, domain, and SSL needed
Before going public: generate real secret keys, add a real domain, and get an SSL certificate.
Check the live site
Enter the API address to see if everything is running. Leave it empty to check this page.
Example: https://api-kappa-pied-59.vercel.app