Hume Octave takes 5,000 characters per utterance in MP3, WAV or PCM

Hume's text-to-speech overview lists "5,000 characters per Utterance" and delivery as MP3, WAV or PCM. Sume TTS 1.0 accepts 1 to 20,000 characters per request and an output_format with an mp3, wav or raw container. The three names line up; the limit is four times larger on Sume.

Hume's figures are from its overview, read 2026-10-01; Sume's from the OpenAPI schema behind the API reference.

How do the specs line up?

Same formats, different limits and delivery.

How do I get PCM on Sume?

Use the raw container with a PCM encoding. The schema offers pcm_f32le, pcm_s16le, pcm_mulaw and pcm_alaw for wav and raw, and sample rates of 8000, 16000, 22050, 24000, 44100 or 48000. The default is mp3 at 44100 Hz and 128 kbps.

curl -X POST https://api.sume.com/v1/tts-1.0/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "transcript": "Raw samples, please.",
    "avatar_handle": "@narrator",
    "output_format": { "container": "raw", "sample_rate": 24000, "encoding": "pcm_s16le" }
  }'

Does an utterance on Hume equal a request on Sume?

Not necessarily. Hume's page says utterance, and the overview does not define whether a request can hold several. Sume's limit is per request and per transcript. Treat each as its own unit and test with your longest script before relying on either number.

Can I stream audio from Sume as it is generated?

No. Phase 1 is an async job with poll or webhook, non-streaming, per the API reference. For live playback, use a streaming engine; for rendered video and audio files, the job result is what you want.