One text-to-speech platform, every voice workload.

Gandr is one speech engine behind every voice workload — long-form narration, dubbing, e-learning, games, podcasts, and real-time agents — priced as a flat, unlimited seat instead of a per-character meter. The engine does not change with the job, and the bill does not change with the volume.

$150

a seat — unlimited speech, any workload

One engine, not a phone box

Real-time voice agents are one thing Gandr does well, but they are one workload, not the product. The same endpoint that voices a live agent narrates an audiobook, dubs a film into 23 languages, reads a course module, and voices a game — because it is a general text-to-speech engine, not a telephony feature. What ships to a call centre and what ships to a studio come out of the same request.

Why a flat seat is hard to pass up

Every other speech vendor prices by the character or the minute: the more you say, the more you pay, and your best month is your most expensive one. Gandr prices by the concurrent seat — $150 a month, unlimited speech on it. A chapter costs what a sentence costs; a busy quarter costs what a quiet one does. Past roughly 5,000 spoken minutes a month a seat is already cheaper than a meter, and it keeps that price while the meter climbs with your success.

The same seat, every workload

Fig

$0 / char

characters, unmetered

$0 / min

minutes, unmetered

languages, one engine

107 ms

first audio, measured

One flat seat covers narration, dubbing, e-learning, games, podcasts and agents alike — the workload changes, the price does not.

The whole platform on one request schema

Streaming or one-shot, cloned voice or stock, any of 23 languages — every capability is a field on one request, over REST, SSE, or a WebSocket. There is no separate product to buy for narration versus agents, no per-voice fee, and no rate-limiting code to protect your own wallet, because the bill is flat.

Notes — an engineer's checklist

01Is Gandr a text-to-speech platform or a voice-agent API?

A text-to-speech platform. Voice agents are one workload it serves, alongside audiobooks, dubbing, e-learning, games and podcasts — all on one engine and one flat seat.

02What does the flat seat actually cover?

Unlimited speech on one concurrent stream, $150 a month: any number of characters or minutes, any of the 23 languages, stock or cloned voices, streaming or one-shot. Concurrency is what you add, not talk time.

03When is a metered vendor cheaper?

Below about 5,000 spoken minutes a month a per-character meter can cost less, and we say so. The seat wins on volume, and on the certainty of a bill that does not move.