Skip to content
Updated for 2026

About 15-20 spoken words — roughly 2.5 words per second at a natural, unhurried pace. Push past that and the clip doesn't get longer; the model compresses the delivery instead, so the character speed-talks through the line and it sounds rushed. Count only what's inside quotation marks, and cut dialogue, not scene description, to fit.

How Much Dialogue Fits in an 8-Second AI Video?

By MakeThisVid Editorial Updated Methodology

Every 8-second AI video clip has a fixed time budget, and dialogue has to fit entirely inside it. Write more than about 15-20 words of spoken dialogue and the model doesn't add time to say it — it compresses the delivery to make it fit. The character talks faster, the pacing feels off, and every word is technically present but the line sounds rushed and unnatural. This is the single most common complaint about AI video dialogue, and almost nobody realizes the cause is word count rather than a rendering fault.

Make a video from one prompt. $2.99 a video · Ready in about a minute · No subscription.

Key facts

Practical word ceiling 15-20 words For an 8-second clip at natural delivery
Speaking rate ~2.5 words/sec With room to breathe, not rushed
What counts Quoted speech only Scene description doesn't compete for time
Over the limit Compression, not extension The clip stays 8 seconds; delivery speeds up
Clip length 8 seconds Fixed — dialogue must fit inside it

How to Use MakeThisVid

From prompt to downloadable MP4, ready to deploy.

  1. Count only the words inside quotation marks

    Scene description does not count against the dialogue budget — only what the character actually says aloud. "A weathered fisherman on a wooden dock at dawn says, 'The tide's turning, we should go now.'" has a long scene description but only 8 spoken words. That fits comfortably.

  2. If you're over ~20 words, cut the line, not the description

    Dialogue is what gets compressed to fit the clip; the visual description does not compete for the same time budget. Trim the sentence itself rather than shortening the scene setup around it.

  3. Split a long exchange across two clips

    If the conversation genuinely needs more than one exchange, generate it as two separate 8-second clips rather than cramming both lines into one. Each clip gets its own full word budget.

  4. Add a pacing instruction at the margin

    A phrase like "speaking slowly and calmly" can help when a line is close to the ceiling. It cannot rescue a line that is simply too long — be honest about that limit before you rely on it.

  5. Consider no dialogue at all

    Silence is a legitimate choice. A held look with no line often reads better on screen than four rushed sentences crammed into 8 seconds.

Make a video — $2.99

Who Uses MakeThisVid for This

Product ad voiceover lines

A single punchy line — a tagline or one short claim — fits the 8-second budget cleanly and reads as intentional rather than rushed.

Character-driven story beats

One exchange of dialogue per clip, split across multiple clips for a longer scene, keeps each character's delivery natural instead of compressed.

Testimonial-style clips

Short, quotable lines under 20 words read as a real person speaking. Longer scripted paragraphs come out sounding like an ad read at double speed.

Frequently Asked Questions

The script is over the word budget for an 8-second clip — roughly 15-20 words at natural delivery. The clip length is fixed, so the model compresses the speech to fit the full line into the available time instead of cutting words or extending the clip.
No. Clip length is fixed at 8 seconds regardless of how much dialogue is in the prompt. Writing more words does not extend the render — it forces faster delivery of the words you wrote.
No. Only the words inside quotation marks — what the character actually says aloud — count against the ~15-20 word ceiling. A prompt can have a long, detailed scene description and still leave plenty of room for dialogue, as long as the spoken line itself stays short.
Partially. A pacing instruction like "speaking slowly and calmly" helps at the margin, when a line is close to the word ceiling. It cannot make a genuinely too-long script sound natural in 8 seconds — the fix for that is cutting words, not adding a pacing note.
Split it across multiple 8-second clips rather than fitting both lines into one. Each clip gets its own full dialogue budget, so both characters can speak at a natural pace instead of both being compressed to fit a single clip.
Often, yes. A silent, held shot reads as more intentional and cinematic than a line that sounds sped up. If a script can't fit in 15-20 words without losing its meaning, cutting the dialogue entirely is a legitimate choice.

Write dialogue that fits the clip

Keep spoken lines to 15-20 words and every 8-second video comes out sounding natural, not rushed.

Make a video — $2.99

Broken renders are remade or refunded.

or 3 more takes — $6.99 Any photo, any look · About $2.33 a take · No subscription required.