Deepgram Flux TTS speed 0.5 to 1.5 vs Sume's 0.6 to 1.5

Deepgram's Flux TTS speed now runs from 0.5 to 1.5 in 0.05 increments, up from 0.85-1.15, and the default stays 1.0. Sume TTS 1.0 takes generation_config.speed as a multiplier in [0.6, 1.5]. The lower bound differs by 0.1, and Sume's docs do not state a step size.

Deepgram facts are from its changelog entry of 2026-08-31, read 2026-10-01. Sume facts are from the OpenAPI schema (API reference).

What changed in Deepgram's range?

Per the changelog, every previously accepted value still works and the wider range applies wherever speed is accepted on Flux TTS: the /v2/speak WebSocket, the batch REST endpoint and Voice Agent sessions on provider version v2. A value outside the range is rejected with SPEED_OUT_OF_RANGE, and one off the 0.05 step with SPEED_INCREMENT_INVALID.

What are Sume's speed, volume and emotion fields?

Inside generation_config: speed in [0.6, 1.5], volume in [0.5, 2.0], and an optional emotion guide string. There is also a top-level speed enum of slow, normal and fast, marked deprecated: prefer generation_config.speed.

How do I port a speed value?

Clamp to your target range before sending, and round to 0.05 if you target Deepgram. A value of 0.55 works on Deepgram but falls below Sume's minimum, so clamp it to 0.6. For the other controls, see text-to-speech with emotion, speed and volume.

function clampSpeed(speed, min, max) {
  return Math.min(max, Math.max(min, speed));
}

const sumeBody = {
  generation_config: { speed: clampSpeed(0.55, 0.6, 1.5) },
};
console.log(sumeBody);