hume/

Hume Octave-2

Public model reference for hume/octave-2: Speechbase pricing, BYOK fee, timestamp behavior, language coverage, and quick-start code.

hume/octave-2

Quickstart Provider docs

Pricing

Managed routing

Speech only: $282 / 1M chars
With timestamps: $282 / 1M chars

Speechbase pays the provider; you spend credits.

BYOK

Platform fee: 3.4%
Speech only: 9.59 cr / 1K chars

Your provider key pays upstream; Speechbase charges the platform fee.

This model returns native word-level timing, so timestamps do not change the managed routing price.

Capabilities

StreamingSupported

Audio tagsNot supported

Voice cloningNot supported

Open sourceNot supported

Quick start

import { writeFileSync } from "node:fs";
import { generateSpeech } from "@speech-sdk/core";

const result = await generateSpeech({
  model: "hume/octave-2",
  voice: "your-provider-voice-id",
  text: "Hello from Speechbase.",
  output: { format: "mp3" },
});

writeFileSync("hello.mp3", result.audio.uint8Array);

curl -X POST https://api.speechbase.ai/v1/audio/speech \
  -H "Authorization: Bearer $SPEECHBASE_API_KEY" \
  -H "Content-Type: application/json" \
  --output hello.mp3 \
  -d '{
    "model": "hume/octave-2",
    "voice": "your-provider-voice-id",
    "text": "Hello from Speechbase.",
    "output": "mp3"
  }'