30-day availability
99.97%
Worst day 99.66%
TexAPI — All systems operational
Refreshes every 60 secondsNo open incidentsMeasured
Mean probe latency
23 ms
Last 30 days
Requests, last 24 hours
138
Peak 12 req/h
Service success rate, last 24 hours
100%
0 service errors
Service status
Each service is probed every 5 minutes; every column below is one day.
- Met
- Minor dip
- Disruption
- Unmeasured
- Last 30 days
The API
The API's internal endpoints
Operational99.99%
API gateway
Routes requests to models
Operational100%
Console
The admin and customer console
Operational99.91%
Last 24 hours of activity
Measured directly from request logs — not simulated.
15:00Peak 12 req/h14:00
Green columns are successful requests; the red cap is failures.
- Organizations active
- 1
- Models called
- 1
Incident history
Last 30 days.
No incidents recorded in the last 30 days.
This page records every disruption that reaches a request automatically, whether or not anyone reports it.
Account situation
IllustrativeEvery request runs on real subscription packages — Claude Max ×20 and ChatGPT Pro — inside a rotating pool: an account near its limit is swapped out immediately, a faulty one quarantines itself.
Only some active accounts are shown — 20 more are rotating inside the pool.
Per-card numbers are illustrative: the magnitudes mirror each channel's real traffic, the individual card values are simulated — not per-account telemetry. Channel quality and rotation counts are real.
How does the pool work?
Multiple accounts run in parallel for each model family. When one approaches its 5-hour rolling limit, requests shift to another before it drains — no interruption, no key change on your side. A faulty account quarantines itself and returns once healthy. That is why there is no such thing as "your own account": every customer shares a pool kept deliberately over-provisioned.
Real subscriptions, not cracked keys
The channels run on properly paid subscriptions — no shared keys, no violating accounts.
Incidents are compensated
Every incident that affects your requests is recorded and compensated under the SLA policy, settled monthly.
Rotating pool, self-quarantine
Many accounts run in parallel: a draining quota hands over early, a faulty account leaves the rotation on its own.
Credit never expires
Top up once, use it whenever — wallet balance carries no expiry date and no maintenance fee.
About this page
Monitoring numbers refresh continuously from the system itself. Upstream channels rotate, quarantine themselves when faulty, and fail over before their quota drains — so you never have to wonder "is it down today?".
The system probes itself, so a total outage (the whole VPS) is recorded as a gap rather than a failure. This is not an SLA commitment; see the SLA policy for the contractual number.
Read the SLA policy →