Pricing · real render costs
What an AI avatar video really costs
On Percify everything runs on one credit meter, and these numbers come from real renders, not a rate card: a talking avatar clip averages 69 credits at 480p and 155 at 720p, a custom avatar from your photo is 5 credits, a voice clone 5, a two person podcast 150 credits a minute, and a 15 second realtime clip about 60. Resolution moves the bill more than anything else. There are no quality, speed or watermark tiers: paid plans add credits, and the free plan starts you with 10 and no card.
- 69
- credits, average 480p talking clip
- 155
- credits, average 720p talking clip
- 5 + 5
- credits to clone a face and a voice
- 10
- free credits, no card needed
The price list, from real renders
Every action draws on the same meter, shared across images, voices and video. There is no separate audio subscription and no per minute voice bill on top.
| Action | Credits | Notes |
|---|---|---|
| Custom avatar from one photo | 5 | Once |
| Generate from your avatar | 2 | Per image |
| Voice clone | 5 | Once, from a sample of 5.6 s or more |
| Short audio clip | 3 | A line of narration or dialogue |
| Longer audio clip | 5 | |
| Talking avatar, 480p | 69 | Average of 121 renders |
| Talking avatar, 720p | 155 | Average of 52 renders |
| Two person podcast | 150 a minute | 15 s to 3 min, 480p |
| Realtime video, 15 s at 480P | 60 | Clips of 5, 10 or 15 s |
Resolution is the whole bill
A 720p talking clip averages about 2.2 times the credits of a 480p one for the same script. By engine setting, moving from the fast 480p path to 720p multiplies credits by 2.4 and render time by 3.3.
What 720p buys is more pixels on the mouth, not better sync. It is worth paying for when the face is small in frame or the clip will be watched full screen; for a face that fills a phone screen, it is money spent on the background.
| Engine setting | Renders | Render seconds per audio second | Average credits |
|---|---|---|---|
| InfiniteTalk fast, 480p | 107 | 7.62 | 65.9 |
| InfiniteTalk, 720p | 52 | 25.15 | 155.2 |
| InfiniteTalk, 480p | 14 | 11.23 | 93.4 |
What a finished video usually costs
The median clip people keep on Percify is 19 seconds, so the typical talking avatar video is one render at 480p: about 69 credits. Add 10 credits once if it should be your own face and voice. A two minute case study at 720p is a very different bill, and only 2.8% of finished clips run that long.
A podcast is priced by the minute instead: 150 credits for one minute, 450 at the three minute cap.
Time is a cost too
On the fast 480p path the engine spends about 7.62 seconds of compute per second of audio, so a 30 second clip is ready in about 3 minutes 49 seconds. The same clip at 720p takes about 12 minutes 34 seconds. When you are still changing the script, that is the difference between iterating and waiting.
How to spend less
Most of the savings are decisions made before you press generate:
- Render at 480p for social. Two thirds of jobs do, and keep the result.
- Write for twenty seconds, and split longer scripts into two clips.
- Crop close before generating. About 9% of source images get run twice, almost always for a problem a crop would have fixed. See fix AI lip sync.
- Draft with a preset voice, switch to your clone for the final.
- Clone once. Ten credits of setup makes every later video yours at no extra cost.
What paid plans change
Only credits, plus API access from the Scale plan up. There is no watermark on any plan, no priority queue and no resolution locked behind a tier. Current plans and their monthly credits are on the pricing page.
Estimate a project in three steps
- 1
Estimate the length
Most clips that get published are around 20 seconds. Count one render per clip.
- 2
Pick the resolution
About 69 credits at 480p, about 155 at 720p, per talking clip on average.
- 3
Add setup once
5 credits for your face and 5 for your voice, if the video should be you.
What those credits buy
Real Percify output at the settings people actually use.
Product ad
a short spokesperson clip
Podcast
a podcast clip, no camera
Explainer
a walkthrough introduction
Frequently asked questions
How much does an AI avatar video cost?
On Percify a talking avatar clip averages 69 credits at 480p and 155 at 720p, measured across real renders. A custom avatar from your photo is 5 credits and a voice clone 5, both once.
Why is 720p so much more expensive?
It costs about 2.2 times the credits and takes over three times as long to render. It puts more pixels on the mouth but does not improve the sync itself, so it only pays off for small faces or full screen viewing.
Is there a free plan?
Yes. The free plan has 10 credits and needs no card, which covers cloning your face and your voice.
Do paid plans unlock better quality or remove a watermark?
No. Paid plans add credits, plus API access from the Scale plan up. There is no watermark on any plan and no quality or speed tier.
How much does an AI podcast cost?
A two person video podcast is 150 credits per minute, from 15 seconds up to a 3 minute cap, at 480p.
Keep reading
Start with ten free credits
Clone your face and your voice on the free plan, then render when the script is ready.