AI agents are getting autonomous. They need to know when to act, what they're allowed to do, and when to ask permission. Simple APIs that keep humans in the loop.

Each tool handles one boundary. Together, they keep your agent responsible and your humans in control.
Know when humans are available
Your agent checks availability before reaching out. Calendar sync, custom states, webhook notifications.
curl https://agentz.fyi/api/v1/check \
-H "Authorization: Bearer az_xxx"
# → { "available": true, "state": "available" }Make your agent ask first
Define rules for what your agent can do autonomously vs. what needs approval. One API call before every action.
curl -X POST https://agentz.fyi/api/v1/check-action \
-H "Authorization: Bearer cg_xxx" \
-d '{"action": "send_email", "category": "email"}'
# → { "allowed": true, "decision": "allow" }More tools coming soon
Escalation routing, action budgets, trust levels...
Use multiple products in a single agent workflow.
// Before your agent acts, check availability + consent
// 1. Is the human available?
const status = await fetch("https://agentz.fyi/api/v1/check", {
headers: { "Authorization": "Bearer az_xxx" }
}).then(r => r.json())
if (!status.available) {
console.log("Human is busy — queuing action for later")
return
}
// 2. Is this action allowed?
const consent = await fetch("https://agentz.fyi/api/v1/check-action", {
method: "POST",
headers: {
"Authorization": "Bearer cg_xxx",
"Content-Type": "application/json"
},
body: JSON.stringify({
action: "send_email",
category: "email",
metadata: { recipient: "boss@company.com" }
})
}).then(r => r.json())
if (consent.decision === "deny") {
console.log("Action blocked by user rules")
return
}
if (consent.decision === "ask") {
console.log("Waiting for human approval...")
return
}
// 3. All clear — execute the action
await sendEmail("boss@company.com", subject, body)One subscription. Every product. No per-API billing.
or $90/year (save 17%)
Start free. Add one API call. Your agent just learned boundaries.
Get Started Free →