OneProvider

Live API Status and Uptime

Is OneProvider up right now? See live API uptime, request volume, the availability of each model family and a clear log of recent incidents.

System status Last sampled 2026-10-17 10:16 UTC. Refreshed from internal probes every 3 minutes; component tiles below reflect the latest sample. Subscribe via Telegram EU Frankfurt probe live PENDING API uptime, 30d API requests, 24h PENDING Components

Model family availability

Checking model observations… Recent incidents

No incidents in the last 30 days

When something breaks we log it here with date, affected system, and resolution time. Telegram channel mirrors the same announcements. All clear. Last probed 2026-10-17 10:16 UTC. Methodology

How we measure

A single Frankfurt probe runs every 3 minutes. While the models answer normally it makes one real model call every 20 minutes. When a model is down it calls on every tick to catch recovery fast. Probe interval (up) 3 min sample · 20 min real Probe interval (down) 3 min real (fast recovery) Window Last 30 days rolling Region EU Frankfurt, single probe Source Live model checks Public history Incident log on request Honest disclosure We do not publish a signed SLA today. Availability is best-effort. Real-time incidents are announced on Telegram — subscribe via @oneprovider_robot for outage notifications.

Frequently asked questions

How is the status measured?

A single probe in Frankfurt runs every 3 minutes. While a model answers normally it makes one real model call every 20 minutes; when a model is down it calls on every tick to catch recovery quickly.

Is there an SLA?

Not a signed one today. Availability is best-effort, and the 30-day uptime figure is a rolling window.

Where are incidents announced?

Here, in the incident log, and in Telegram: subscribe through @oneprovider_robot for outage notifications.

The page is green but my requests fail. What now?

The probe checks one region and one request type, so first check your own side: the status code in error codes , a balance or key limit (429 “Insufficient balance”), and any VPN or VLESS proxy. Then message support with the status code and the exact error text.

Should I retry on a 5xx error?

Yes. 500, 502, 503 and 504 are temporary: retry with a growing pause. Do not retry 401, 403, 404 or a 429 with insufficient balance.

Can I connect it to my own monitoring?

Ask us on Telegram: we can plug a custom health hook into your Slack channel or Grafana board.

OneProvider · API gateway

Want it wired into your monitor?

Ping us on Telegram — we'll plug a custom health hook into your procurement pipeline, Slack channel, or Grafana board. Same-day setup during CET hours.

Get your key Explore docs