A 4-minute explainer from 3,420 characters: $0.16 voice, $0.40 render
Budget a 4-minute explainer on Sume with a 3,420-character script: TTS $0.16, Music 1.0 bed $0.125, Timeline render $0.40. Total about $0.69 before visuals.

A 4-minute explainer built from a 3,420-character script costs about $0.69 on Sume before any visuals: text-to-speech at $0.0475 per 1,000 characters is $0.16, a Music 1.0 bed is $0.125, and a Timeline 1.0 render of 4 billed minutes is $0.40. The visuals, whether stills or clips, are the real budget.
Script to minutes
As a planning assumption, narration runs at about 15 characters per second, so 3,420 characters is 228 seconds, or 3 minutes 48 seconds. That is a rule of thumb, not a Sume guarantee; the real length is the length of the TTS audio, which you measure when the job returns. Sume's TTS has a maximum request of 20,000 characters, so this script is a single request.
Render billing follows the output length rounded up: 228 seconds is 4 billed minutes. If you trim the script by 200 characters, the narration is about 215 seconds, still 4 minutes. The line that moves the bill is the next minute boundary, 180 seconds, which is 2,700 characters.
| Line | Arithmetic | Cost |
|---|---|---|
| TTS narration | 3,420 / 1,000 x $0.0475 = $0.16245 | $0.16 |
| Music 1.0 bed | one generation, flat | $0.125 |
| Timeline 1.0 render | ceil(228 / 60) = 4 x $0.10 | $0.40 |
| Total | $0.69 |
Putting it together
The narration is the audio spine: set audio.url to the TTS file and audio.duration_seconds to its measured length. Use soundtrack for the music bed with a duck_db value between 0 and 20 so the bed drops under the voice. Ducking needs a real spine, which the narration is.
If you generate narration in several parts to review each, join them with Timeline audio concat, a sample-domain join with no re-synthesis, for $0.01 per job, or pass audio.parts to the render with no extra job.
The part that is not cheap
Visuals decide the price. Eight stills at nano-banana-2.1 1K are 8 times $0.10, which is $0.80, and clips cost per second by model. Add that to the $0.69 only after you have chosen a style. Price the plan with the free Timeline plan call before you spend on anything generative, and check live rates in the catalog.
A practical note
Review the narration before you render. TTS reads exactly what you give it, so spell out numbers, expand abbreviations and add punctuation where you want a pause. Generate the voice first and measure its length, then cut or add visuals to match, because the spine sets the output length. A script change after the render costs a new $0.40 render, while a change before it costs nothing, so lock the words first.
Before you scale
Every price in this post is a Sume list price read from the public catalog on 2026-10-09, and the arithmetic is shown so you can redo it with your own counts. Rates can change, so re-read the catalog before a large batch and run a small test first. Sume bills the job's captured amount, and the job result tells you what was used, so compare the first run's receipt with your estimate before you scale the work to the full set.
Sources
Related posts
More in Use cases
- 40 ad backgrounds on gpt-image-2.5: $0.99 low, $2.23 mid, $8.90 high
Price 40 ad background variants on Sume's gpt-image-2.5: low 1K $0.99, medium 2K $2.23, high 4K $8.90. Pick the quality tier before you generate.
- 40 die-cut stickers: $0.99 transparent, or $1.90+ with RMBG
Forty stickers cost $0.99 on GPT Image 2.5 low with a transparent background, or $1.90 to $4.90 with another model plus Sume background removal.
- 45-second 9:16 ad for TikTok, LinkedIn and Pinterest: checks and $0.32
A 45-second 720x1280 ad clears TikTok, LinkedIn and Pinterest length and size lists. The Sume pipeline is trim $0.02, timeline $0.10, captions $0.20: $0.32.
- 45-second teaser from a 12-minute podcast on Sume: $0.43
STT to find the quote, split 45 seconds, render over a cover still, burn captions: $0.12 + $0.01 + $0.10 + $0.20 = $0.43. Each step and its limit.
Written by Sume