Support CenterFAQGeneral Questions
Back to FAQ

General Questions

Frequently asked questions about Agent Status and agent validation.

What is Agent Status?

Agent Status is an agent validation platform that monitors your AI agents from real devices around the world. It answers the question: "Is my agent actually working for users right now?"

How is Agent Status different from uptime monitoring?

TraditionalAgent Status
HTTP status checkFull response validation
Single locationMultiple global regions
Yes/no availabilityUP/DEGRADED/DOWN verdicts
No semantic checksGold prompt verification
Fixed test serversReal user devices

What agents does Agent Status support?

Any AI agent with an HTTP API: OpenAI, Anthropic Claude, Google Gemini, Mistral, HuggingFace, LangChain/LlamaIndex, or custom agents.

How often does Agent Status check my agent?

TierMinimum Interval
FreeEvery 6 hours
Hourly ($10/mo)Every 1 hour
Standard ($20/mo)Every 15 minutes
Pro ($50/mo)Every 5 minutes

Where does Agent Status test from?

Validation nodes in US, EU, AP, LATAM, AF, and ME — real devices in homes and offices, not data centers.

What data does Agent Status store?

  • Stores: Timestamps, latencies, error codes, verdicts, pass rates
  • Does NOT store: Your agent's responses, user prompts, or credentials

Is my agent's auth token secure?

Yes — encrypted at rest, never logged, never returned in API responses.

Can Agent Status break my agent?

No. Agent Status only sends read-only chat completion requests. No destructive operations.

How much traffic does Agent Status generate?

Standard tier example (15 min, 10 nodes): ~28,800 requests/month. Negligible for most services.

Does Agent Status support webhooks?

Yes — status change notifications (DOWN, DEGRADED, UP, recovery).

Can I self-host Agent Status?

Not currently. Enterprise dedicated infrastructure available on request.

How do I get support?

Email support@carmel.so or use the dashboard help widget.

Need more help?

Our support team is available to assist you

Contact Support