72,000-character audiobook chapter on Sume TTS: $3.42 in 4 requests

At $0.0475 per 1,000 characters, 72,000 characters is $3.420 before rounding, so $3.42. Because one Sume TTS request accepts at most 20,000 characters, the chapter needs 4 requests: three of 20,000 and one of 12,000.

Splitting the chapter

The 20,000-character maximum comes from the TTS model definition in the Sume API contract; spaces and punctuation count toward it. Split at paragraph boundaries so each piece reads cleanly when you join the audio.

Scaling to a book

A 12-chapter book at that length is 864,000 characters, or $41.04. That is a rate-card figure for the voice track only. Cover art, narration retakes, and any audio mastering are separate.

Worked example and checks

Cost per hour of narration is a useful sanity check. A common reading pace is 150 words per minute, and English text averages around six characters per word including the space, so about 900 characters per minute. That is an estimate rather than a Sume figure, but it implies about $2.57 per hour of finished audio at $0.0475 per 1,000 characters (54,000 characters x 0.0475 / 1,000). Treat it as order of magnitude only.

Revisions are where TTS bills grow. Re-reading one 20,000-character piece costs $0.95 again. Keep a table of which pieces were re-rendered and why; two re-reads of a chapter's four pieces add $6.84 to the $3.42 first pass.

What Sume does not do

Sume joins nothing for you at the TTS level. If you need one audio file, assemble the pieces yourself or use the Timeline route with the pieces as the audio spine. Sume bills model and tool usage at the published rate on top of the plan fee, so the plan price is separate from every number in this post.