Skip to content

who it’s for

verified July 2026

Gandr for enterprise fleets.

An enterprise voice bill on a character meter is a forecast that has to be re-approved every quarter. A fleet of flat lines is a number that survives the fiscal year: lines × $150, whatever your callers say. Every claim on this sheet is published elsewhere on this site with its measurement basis stated.

01

What signs off in one meeting

  • The bill is lines × rate — finance models it in one cell, and doubling call volume changes nothing on the invoice.
  • Latency is published as percentile pairs with the harness stated, and the pair holds on every line in the fleet.
  • Every synthesized sample carries an inaudible watermark, and customer reference audio never enters a training set.
  • Standby failover lines idle beside your current vendor at $20 a month — adoption can start as insurance.

02

A fleet, priced

Every meter cell is minutes × 1,000 characters × the published $30–$100 per-million range. The Gandr column is line count × $150 — check any row by hand.

Table 1 — Three fleet sizes, one month, both models

FleetMinutes / moOn a $30–$100 meterOn Gandr lines
20 lines, 24/7864,000$25,920 – $86,400$3,000
50 lines, business hours540,000$16,200 – $54,000$7,500
100 lines, 24/74,320,000$129,600 – $432,000$15,000

03

The posture behind the price

Three serving tiers stand behind one endpoint: the GPU fleet, a warm spill, and an independent fallback — a full outage costs latency, not availability. Past absolute capacity the API refuses fast with a retryable 503 instead of degrading a live call. The reliability sheet publishes the whole posture; the provenance sheet covers watermarking and data handling.

See also

Related sheets.

Thirty minutes on your own numbers — fleet size, talk time, and the bill both ways, worked live.

Bring your stack to the call