Rime Coda 180 s audio limit vs Sume TTS 1200 s limit

Rime's Coda 1.0.5 changelog says the configured maximum audio duration is 180 seconds, and a unit-rate fix means it no longer cuts a generation off at 172.8 seconds. Sume TTS 1.0 allows far longer output: synthesized audio over 1200 seconds fails with tts_duration_exceeded, with no credit capture. For anything longer, split the script and join the audio.

Rime facts are from its changelog; Sume facts from the OpenAPI reference and Timeline audio docs, read 2026-10-01.

What changed in Rime Coda?

The changelog says Coda bounds long generations at sentence boundaries while preserving audio order within a request, and that its maximum audio duration calculation now uses the codec's correct rate. The configured 180-second limit therefore applies as written, not 172.8 seconds. Plan chunks under 180 seconds on that service.

What are Sume's per-request limits?

How do I produce audio longer than the limit?

Split the transcript at paragraph or sentence boundaries, submit each part as its own TTS job, then join the audio files. Timeline audio's concat operation takes parts[] of 1 to 20 ordered items from your own media.sume.com audio, and the docs describe it as a sample-domain join with no re-synthesis, so import generated files first. With 20 parts of up to 1200 seconds each, that is a lot of headroom for a chapter or a full episode. See TTS chapters past the 1200-second limit.

What should I watch for?

Use the same voice and language on every part so the joined file sounds consistent. Cut at sentence ends, not mid-sentence, and keep the part order with your job ids.