Skip to content
ClaudeOpus

Service

Status

Two live probes of the service, run from this browser. Each readout says what was asked, what came back and when — nothing is averaged, estimated or carried over from another visitor.
Probed from your browserEvery 30 secondsTimes in UTC

Running the first checks…

Probing the Model API and the Account API…

Probes

Model API

Checking…
GET /v1/models

The endpoint your tools call for Claude: /v1/messages and the OpenAI-compatible /v1/chat/completions.

Answer
Response time
Checked
Last as expected

Last 30 checks

None yet

Expects 200 with the model list. While the service is in maintenance, every /v1 route answers 503.

Account API

Checking…
POST /api/key-status

Key lookups, sign-in and the dashboard.

Answer
Response time
Checked
Last as expected

Last 30 checks

None yet

Looks up a key that doesn’t exist and expects 404 “not found” — proof the service and its database both answered.

What these probes can’t see

  • Neither sends a model request — that needs a key and would use its 5-hour window. A fault that only affects model responses won't show here.
  • They run from your browser, so response time includes your own network, and an outage on your side looks the same as one on ours.
  • A probe gets 10 seconds to answer. Its history lives in this tab only: close it and the record is gone.
  • If your requests fail while both probes answer, check the key itself on Key status.

Run the same checks

The exact requests this page sends, for a terminal:

Model API
curl -i https://api.claudeopus.pro/v1/models
Account API
curl -i -X POST \
  https://api.claudeopus.pro/api/key-status \
  -H 'Content-Type: application/json' \
  -d '{"key":"health-check"}'