Gandr for enterprise fleets.
× $150
the whole budget line, headcount arithmetic, not usage forecasting
One engine across the whole estate: narration, dubbing, localisation and live calls on the same endpoint, the same six voices, the same auditable key per workload. Then a budget line procurement can sign once.
What signs off in one meeting
- Your bill is streams × rate. Finance models it in one cell. Double your speech volume and the invoice does not move.
- First audio byte in 146 ms over the open internet, 116 ms p50 first audio, server side warm.
- Every sample we synthesize carries an inaudible watermark. Your reference audio never enters a training set.
A fleet, priced
Every meter cell is minutes × 1,000 characters × the published $30, $100 per-million range. The Gandr column is stream count × $150, check any row by hand.
table 01
Three fleet sizes, one month, both models
| Fleet | Minutes / mo | On a $30, $100 meter | On Gandr streams |
|---|---|---|---|
| 20 streams, 24/7 | 864,000 | $25,920, $86,400 | $3,000 |
| 50 streams, business hours | 540,000 | $16,200, $54,000 | $7,500 |
| 100 streams, 24/7 | 4,320,000 | $129,600, $432,000 | $13,500 |
The posture behind the price
Serving is redundant behind one endpoint, so an outage costs latency, not availability. A request that cannot be served is refused fast and retryably rather than degrading a live call.
A key and one stream to size it on, before anyone talks about a fleet.
Get a key for your stack