OpenAI speech API has six formats; Sume TTS has mp3, wav, raw

OpenAI's text-to-speech guide lists six response formats: MP3, Opus, AAC, FLAC, WAV and PCM, with PCM described as raw 24 kHz 16-bit signed little-endian samples without a header. Sume TTS 1.0 offers three containers: mp3, wav and raw. If your pipeline needs Opus, AAC or FLAC, plan a conversion step on your side.

OpenAI's list is from its guide, read 2026-10-01; Sume's from the OpenAPI schema behind the API reference.

Which formats map to which?

Only three line up directly.

What can I tune on Sume's formats?

Sample rate can be 8000, 16000, 22050, 24000, 44100 or 48000 Hz. MP3 bit rate can be 32000, 64000, 96000, 128000 or 192000. WAV and raw take pcm_f32le, pcm_s16le, pcm_mulaw or pcm_alaw. MP3 cannot carry a PCM encoding.

What do I send to match OpenAI's PCM?

Raw container, 24000 Hz, signed 16-bit little-endian.

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": "Match the other API.",
    "avatar_handle": "@narrator",
    "output_format": { "container": "raw", "sample_rate": 24000, "encoding": "pcm_s16le" }
  }'

Is there an OpenAI-compatible endpoint?

See the OpenAI-compatible audio speech endpoint post for what the Sume route accepts; this post covers the output formats only.