45,000-character script: three Sume TTS jobs plus a join for $2.17

A 45,000-character script needs three TTS requests under the 20,000 cap. At 72 cents each plus a one-cent join the audio costs $2.17 on Sume.

5 min readSume
All posts

The plan and the total

A 45,000-character script exceeds the 20,000-character request cap, so send three 15,000-character jobs, then concatenate them. Each job costs 72 cents; three cost 216 cents; the join adds 1 cent, for 217 cents ($2.17) of audio work.

Sume TTS list price is 38 micro-dollars (0.0038 cents) per character; the billable price is list x 1.25, rounded up to whole cents per job. That is 0.00475 cents per character before rounding, so a 20,000-character request is 20,000 x 0.00475 = 95 cents.

Arithmetic by step

15,000 x 0.00475 = 71.25 cents, rounded up to 72 cents per job. Splitting evenly keeps each part safely under the cap; splitting 20,000 + 20,000 + 5,000 would cost 95 + 95 + 24 = 214 cents, one cent less because of rounding, but leaves you with an uneven last part.

Cost of narrating 45,000 characters on Sume (rates read 2026-10-08)
StepQuantityUnit priceCost
TTS jobs3 x 15,000 characters72 cents$2.16
Concat join (Timeline audio)1 job$0.01$0.01
Total$2.17

For reference, vendor per-character rates

At the tracker's listed rates the same 45,000 characters would be $0.99 on MAI-Voice-2.1 and $0.675 on MAI-Voice-2.1-Flash. Sume does not list MAI-Voice-2.1, MAI-Voice-2.1-Flash or MAI-Transcribe-2 in its catalogs, so this is a price comparison, not a statement that you can call Microsoft's models through Sume. Vendor prices come from a third-party release tracker that lists them, not from a Microsoft price page, so confirm them on Microsoft's own pricing before you commit.

Join details

Timeline audio concatenates Sume-hosted audio in the sample domain with no re-TTS and no silence at the seams, and it returns a durable media.sume.com file with timings. Output is capped at 1,800 seconds, so very long audiobooks need more than one joined file (Timeline audio).

Cut at paragraph ends so the join lands between sentences, and keep the same model id, voice and speed on all three parts.

Where to cut the script

Cut at the ends of paragraphs or scenes, never in the middle of a sentence, because prosody resets at the start of every request. If a section break is natural, such as a chapter title, cut there and the join will sound intentional.

Keep parts roughly equal so a failed part is a cheap retry. Three 15,000-character parts mean the worst single retry is 72 cents. Log the character count of each part so you can reconcile the bill: 15,000 characters bills 72 cents every time.

Checks after the join

Listen to the two seams, then check the total duration against the sum of the part durations; the join returns timings for the pieces. If you plan captions next, caption the joined file once so the timing is continuous. Remember captions are $0.20 per job for clips up to 60 seconds; a long audiobook with captions needs the live price from the catalog rather than this fixed estimate.

Sources

Related posts

More in Use cases

All Use cases posts

Written by Sume