Avatar Studio · talking head video
A talking head video from one photo
An AI talking head generator turns one clear photo and a script into a video of that face speaking, with the mouth driven by the audio. On Percify the median clip people keep is 19 seconds, two thirds run under 30 seconds, a 480p render averages 69 credits, and the photo decides quality more than any setting. No camera, no studio, no filming step.
- 19.0 s
- median clip people keep
- 66%
- of finished clips run under 30 seconds
- 168 s
- median render time
- 69
- credits for an average 480p render
What you give it, and what comes back
Two inputs: a photo of a face and the words you want said. It returns a video of that face saying them, with the mouth driven by the audio rather than animated by hand. There is no rig and no capture session.
The voice is either one of 52 preset voices or your own cloned voice, built from a sample of at least 5.6 seconds. Because the model conditions on the photo, most of the quality is decided before you press generate.
The length people actually publish
Every guide to talking head video recommends a two minute case study. Almost nobody makes one. Across all 179 finished renders from 61 accounts between 7 May and 2 September 2026:
Write for twenty seconds. A spokesperson clip earns its place as a hook, an answer or an intro, and all three are short. If the script will not fit, cut it into two clips rather than rendering one long one.
| Measure | Value |
|---|---|
| Median clip length | 19.0 seconds |
| Under 30 seconds | 66% |
| Over two minutes | 2.8% |
What it costs, and why resolution is the bill
Rendering is priced per second of output, and resolution changes the bill more than anything else you choose. 720p is roughly 2.2 times the price of 480p for the same script. The case for 720p is a close up watched full screen; the case against it is almost everything made for social.
It is one credit meter across images, voices and video. There is no watermark on any plan and no speed tier to buy. The full breakdown is in what an AI avatar video costs.
| Resolution | Renders | Average credits |
|---|---|---|
| 480p | 121 | 69 |
| 720p | 52 | 155 |
How long you will wait
The engine spends about 7.5 seconds of compute per second of finished video. Median render time is 168 seconds, and the slowest tenth take 757 seconds or more. In practice a 20 second clip is about two and a half minutes and a 60 second clip is closer to eight. Queue it and do something else.
Choosing a talking head generator
Talking head tools come in three designs, and the design predicts results better than any feature list. Photo driven tools work from a single image of you. Video driven tools preserve a performance you already filmed. Stock presenters give the most consistent sync, but it is not your face.
Five questions sort them quickly:
- Does it start from your own photo, or only from its presenters?
- What does a 20 second clip cost at the resolution you will publish?
- How long does a render take?
- Does it watermark output on the plan you would buy?
- Can it speak in your own cloned voice?
Make one in four steps
- 1
Pick a photo that fills the frame
The face at least a third of the frame width, nothing across the mouth, even frontal light.
- 2
Write twenty seconds
One line of hook, two or three of substance, one of exit. Read it aloud before rendering.
- 3
Choose the voice
A preset for drafts and faceless videos, your clone when the video is about you.
- 4
Render at 480p for social
Save 720p for a close up that will be watched full screen.
Real talking head output
Each of these started as one photo and a script.
Alex
talks from one photo
Christine
a spokesperson close up
Company update
a leadership message, no filming
Frequently asked questions
What should I look for in an AI talking head generator?
That it starts from your own photo, prices a 20 second clip clearly, renders in minutes rather than hours, does not watermark your output, and can speak in your cloned voice. Percify does all five, with one credit meter and no quality tiers.
How long should a talking head video be?
Short. The median clip people keep on Percify is 19 seconds, 66% run under 30 seconds and only 2.8% run past two minutes. Write for twenty seconds and split longer scripts.
How much does an AI talking head video cost?
On Percify a 480p render averages 69 credits and a 720p render averages 155. The free plan has 10 credits and needs no card, which covers cloning your face and your voice before your first render.
Can I use my own face and voice?
Yes. A custom avatar from one photo costs 5 credits and a voice clone costs 5 credits, both once. The voice sample needs at least 5.6 seconds of clear speech; 15 to 30 seconds is better.
How long does it take to render?
About 7.5 seconds of compute per second of video. The median render takes 168 seconds, so a 20 second clip is ready in roughly two and a half minutes.
Keep reading
Make your first talking head
One photo, twenty seconds of script and a voice. The free plan's 10 credits cover cloning your face and your voice.