A 1,000-word script takes about 6 to 7 minutes to read aloud at 150 words a minute, a common planning rate for English. A 10-minute video needs roughly 1,500 words. Treat this as a starting estimate rather than a timing rule: generate a representative passage in your chosen AI voice, measure it, and plan from that.
How many minutes is my word count?
At 150 words a minute:
| Words | Approximate length |
|---|---|
| 100 | 40 seconds |
| 250 | 1 minute 40 seconds |
| 500 | 3 minutes 20 seconds |
| 750 | 5 minutes |
| 1,000 | 6 minutes 40 seconds |
| 1,500 | 10 minutes |
| 2,000 | 13 minutes 20 seconds |
| 3,000 | 20 minutes |
| 4,500 | 30 minutes |
| 10,000 | about 1 hour 7 minutes |
| 80,000 | about 8 hours 53 minutes |
The formula is simple: minutes = words ÷ 150.
How many words do I need for a set length?
Work backwards: words = minutes × 150.
| Target length | Words at 150 a minute |
|---|---|
| 30-second ad | 75 |
| 60-second short | 150 |
| 5-minute explainer | 750 |
| 10-minute video | 1,500 |
| 15-minute episode | 2,250 |
| 20-minute episode | 3,000 |
| 30-minute episode | 4,500 |
For a video with screen action or b-roll, write fewer words than the table says. Moments where the picture carries the story need silence, not narration.
Why is 150 words a minute only an estimate?
Because the real pace depends on things a word count can't see. Some voices simply read faster or slower than others. The writing matters too: short sentences with full stops give more natural breaks than one long run-on sentence.
Numbers and symbols change the count. "$12.50" is one word on the page and "twelve dollars and fifty cents" is five when spoken, so writing numbers out, as you should for a voice, raises the total. See How to write numbers, dates and acronyms for AI voices. Every deliberate pause adds time on top of the words.
Then there's language. 150 words a minute is an English planning figure. Word counts don't compare well across languages, and they're especially unsuitable for Chinese and Japanese, which don't separate words with spaces.
How do I measure the real length?
Measure a sample instead of trusting the formula:
- Pick a representative passage of 200 to 400 words, with the same mix of dialogue, numbers and pauses as the full script.
- Generate it in the voice you'll use for the final.
- Read the length of the generated audio on the timeline.
- Work out your rate: words ÷ minutes. For example, 400 words in 3 minutes is about 133 words a minute.
- Plan the full script with your rate, not the generic one.
If you change the voice later, measure again.
How do I hit an exact length?
Trim words first. Cutting a sentence changes the length more predictably than trying to speed up the read.
Where timing matters, use exact pauses. In Vois Studio, a pause pill adds a silence of a set length, from 0.3 to 3 seconds, so you can hold for a title card or a screen change. See Pause nodes for precise silence.
Don't pad to fill time. Extra filler sounds like filler. Retime the edit to the real audio instead of forcing the audio to a guessed runtime. And for long projects, work in sections, measuring each chapter or segment as you go: Long script chunking strategy.
Related
- How to write a script for AI voices
- Pause nodes for precise silence
- How to write numbers, dates and acronyms for AI voices
- Long script chunking strategy
Generate 300 words in your final voice today, time them, and plan the rest of the script from that number.
The Vois Team