Operations and security

Keep AI running, and stop it in seconds when you must.

Switches that act at once, from a single key to every AI call. Fallback that moves calls before anyone sees an error. Limits that stop one caller starving the rest, and a record of every change that can't be quietly altered.

Vidai Control room · Keys
Suspending a leaked key from its page, with the reason recorded.

Why it matters

Every app finds out at once.

When a provider rate-limits or goes down, every app hits it together, each with its own retry loop. In an incident minutes matter: a leaked key, a runaway agent, a model saying what it mustn't. And nobody can say who changed which setting, when.

What it does

Stop switches at every level

Suspend one key, agent, team, provider or model. Or stop every AI call on every Vidai server, and resume exactly as before.

Fallback and high availability

When a model fails, the next one answers before the caller sees an error. Run several Vidai servers behind your load balancer, with autoscaling, so no single server is a point of failure.

Provider keys that rotate

Several keys per provider; a failing key is taken out of service and the others carry its calls. Rotate with no outage.

Rate limits on every key

So one runaway caller can't starve everyone else, with a live reading of how close each key is.

A tamper-evident audit log

Who changed what, from where and why, people and scripts alike, in records that can't be quietly altered.

Events to your SIEM

Blocks, refusals and changes sent as they happen, signed, never containing prompts or responses.

In the Control room

See it working.

Settings · Stop AI traffic

One switch for every AI call.

Calls already running finish; new ones are refused in each caller's own format. Everyone sees why, and every server confirms it took.

Vidai Control room · Settings
Stopping all AI traffic, with the reason everyone will see.
Stopping all AI traffic, with the reason everyone will see.

Providers · Health

Failing models rested, calls moved on.

Each model's health, when a failing one was rested, and when it's tried again.

Vidai Control room · Providers
Provider health: models in service, rested, and being tried again.
Provider health: models in service, rested, and being tried again.

Routing · Experiments

Try a new model on part of the traffic.

Split calls between two models and compare cost, speed and errors before you move everyone.

Vidai Control room · Routing
An A/B experiment's results, model against model.
An A/B experiment's results, model against model.

Proof · Audit log

Every change, on the record.

People and scripts alike, with where they acted from and the reason they gave.

Vidai Control room · Audit log
The audit log: who changed what, from where and why.
The audit log: who changed what, from where and why.

Who uses it

Platform and SRE teams keep AI running through outages. Security stops what must stop, in seconds. Everyone can see what changed, and who changed it.

Stop switches, routing, fallback, rate limits and the tamper-evident audit log are in every edition. Signed events to your SIEM and single sign-on come with Enterprise.

Related

How we start together

Start with one team, on your own AI traffic.

After a 20-minute demo, we agree a two-week pilot with you: one app or team, your traffic, your infrastructure.
  1. 01

    Pick one app or team

    Point it at the Control room. One configuration change; nothing else changes.

  2. 02

    Watch for two weeks

    Every call priced, checked and recorded. Rules start in observe mode.

  3. 03

    See the evidence

    What it cost, what it caught and what it would have stopped, from your own traffic.