Gateway

holaProxy

One API for every model — unified credits, token metering, and cost control.

Agents get expensive quietly. A model call here, an integration call there, across a dozen workspaces — and by the time it shows up, nobody can say which client it belonged to. holaProxy makes the meter part of the infrastructure instead of an afterthought.

Capabilities

The gateway every agent runs through

One API, every model

Brokered access to high-performance models, plus your own providers behind the same interface. Switching model is a string, not a migration.

Unified credits

Model calls and integration calls draw from one balance. No separate invoices to reconcile at the end of the month.

Token metering

Input, output and cache hits are metered per workspace, so consumption is attributable rather than aggregate.

Cost control

Budgets and caps from department down to member, with the numbers visible before the bill arrives.

How it meters

One balance, one number to watch

1 credit ≈ 2,000 tokens through brokered models. Integration calls draw a small per-call charge, tiered for heavy monthly use. Credits never expire.

See what a pack costs
  1. Step 1

    Bring your keys, or don't

    On your own keys, model calls cost zero credits. Without them, we broker the call.

  2. Step 2

    Every call is metered

    Model tokens and integration calls land on the same meter, tagged to a workspace.

  3. Step 3

    Spend stays legible

    You can answer 'what did this client cost us in agent time' without an export.

Gateway

Stop guessing what agents cost

Point your agents at holaProxy and every model call, integration call and workspace lands on one meter.