Customer stories, produced without a film crew

This AI testimonial video generator turns a topic into a spoken piece delivered by an avatar you choose, across one to three scenes. Each scene sets its own location and action, the lines are written for you unless you supply them, and the finished video runs 5 to 30 seconds.

How it works

  1. 1

    Pick an avatar and topic

    An avatar is required — it fronts every scene. The topic can be a sentence.

  2. 2

    Describe the scenes

    One to three, each 5 or 10 seconds, with its own setting and action.

  3. 3

    Generate

    Lines are written per scene unless you supply your own, then lip-synced.

What you can make

Brands illustrating a use case

Show how a product fits into someone’s day when you have no customer footage to work with.

Pre-launch products

Before you have customers you have no testimonials. An openly-illustrative video fills the gap without inventing a person who exists.

Ad creative teams

Produce several spoken angles quickly to find which message lands, before commissioning real footage.

Localised campaigns

Re-record the same message with a different avatar and voice per market instead of re-shooting.

This AI testimonial video generator turns a topic into a spoken piece delivered by an avatar you choose, across one to three scenes.

Make your first testimonial free

Specs & limits

RequiredAn avatar, and a topic of 3–500 characters
Scenes1 to 3, each with its own setting
Scene length5 or 10 seconds each → 5–30 seconds total
Spoken linesAuto-written per scene, or supply your own
Aspect ratios9:16, 16:9 or 1:1
CaptionsOptional, burned in, top / centre / bottom

Tips for better results

  • Write the scene as a place and an action, not a mood. "Holding the jar at a kitchen counter, morning light" gives the model something to build; "authentic and relatable" does not.
  • Supply the spoken line yourself when a specific claim matters. Auto-written copy is fine for tone, but you should control anything a customer could hold you to.
  • Two scenes usually beats three. A change of setting adds credibility; a third often just adds five seconds people do not watch.
  • Keep one avatar per campaign. Testimonials read as a person, and swapping faces between videos breaks the only continuity this format has.

Frequently asked questions

Is this a real customer testimonial?

No. It is a scripted video delivered by a synthetic presenter — useful for illustrating a use case, not for standing in as a real person’s account. Presenting it as a genuine customer testimonial is deceptive, breaches ad-platform policy, and in most markets falls under consumer-protection and advertising rules. Label it as an illustration where context could mislead.

Do I need an avatar first?

Yes — an avatar is required, since it fronts every scene and is lip-synced to the script.

How long is the video?

Between 5 and 30 seconds: one to three scenes, each 5 or 10 seconds.

Can I write the script myself?

Yes. Each scene takes an optional spoken line; leave it blank and one is written for you.

Is it free?

Generating uses credits from your Publiz plan. Video costs more than image tools because each scene is generated and lip-synced; new accounts start with a free credit allowance.

Related tools

Browse all AI tools