50-minute show on Timeline: two renders, $5.00 if cut on a minute

A 50-minute show cannot be one Timeline render on Sume, because audio.duration_seconds is capped at 1,800 seconds (30 minutes). Split it into two renders. A 30-minute plus 20-minute split costs 30 x $0.10 + 20 x $0.10 = $5.00. A split at 29 minutes 30 seconds gives 30 + 21 minutes and costs $5.10, because each render rounds up on its own.

Why where you cut matters

Timeline rates each render by ceil(output minute) at $0.10. Two renders round separately. If the cut falls exactly on a minute, the total equals the whole show in minutes. If the cut falls mid-minute, both halves may round up, and you pay for an extra minute.

Make the cut at a natural point

Cut at a topic change that falls near a whole minute. If the nearest topic change is at 24:40, either trim 40 seconds or accept one extra billed minute ($0.10). You rarely need to optimize this; the whole point is that the worst case is a dime.

Keep both halves on the same narration engine and voice. A voice recorded by sonic-3.6 in part 1 and a different engine in part 2 will be audible at the join between files.

Join audio without a render

If you want one long audio file instead of a video, Timeline audio concat joins up to 20 Sume-hosted audio parts into one gapless file. The Timeline audio doc does not state a price on that page; check GET /v1/catalog for the rate before you plan a budget around it.

For video, each render is separate, so the show's second half can use a different bed or its own fade-out. Set fade_out_seconds on the final render only.

Quick formula

Total cost = ceil(part 1 seconds / 60) x $0.10 + ceil(part 2 seconds / 60) x $0.10. If both parts are whole minutes, it equals total minutes x $0.10. Each extra partial minute adds a dime. Add narration (characters x 47.5 micro-dollars, per request up to 20,000 characters) and music ($0.125) to get the show's full cost.

A 50-minute show at the planning figure of 900 characters a minute is 45,000 characters, which is the chapter size in the related TTS post.