About

About AvatarSay

AvatarSay (avatarsay.com) is an online tool that turns a text description into a fictional character and a short talking video. It is built for people who want a distinctive face without uploading their own photo.

AvatarSay is not affiliated with other projects of the same name, such as an open-source command-line tool on GitHub.

Last updated: November 7, 2026

Who runs it

An independent product

AvatarSay is built and run by the AvatarSay team — the same team writes the code, answers support and stands behind the policies on this site.

Operator: AvatarSay — an independent product, not part of a larger group
Contact: support@avatarsay.com · replies within 1 business day
Payments: Handled by Waffo, the merchant of record for every purchase. Waffo takes the card details; we never see them.
Official channels: YouTube (@AvatarSayAI) · TikTok (@avatarsayai) · Instagram (@avatarsayai) · Open benchmarks (GitHub)

The same details are set out in our Terms of Service and Privacy Policy.

Key facts

The product in numbers

Every number below is generated from the same price list checkout uses, so it matches the pricing page and the product. Last checked 2026-11-12.

What it makes: An original fictional character from a written description, then a short talking video of that character.
Input: Text only. No photo upload, no video upload, no face swap, no real people.
Images per generation: 2 × 1024 px in one of 8 styles
Video length: 5–30 seconds per paid clip, 12 credits per second
Output: MP4 (H.264 + AAC), 960×960 square, 30 fps, no watermark
Script length: Up to 400 characters
Voices: 8 built-in text-to-speech voices, English only. No voice cloning.
Languages: English only. No translation.
Render time: Short 3–5 s clips usually take 3–5 minutes; 15–30 s clips take about 9–14 minutes.
Price: $9.90 for 500 credits; $24.90 for 1,500 credits; $59.00 for 4,000 credits (USD, one time). Credits never expire.
Free: 30 credits on sign-up plus one free talking clip up to 5 s (75 chars) while the launch offer lasts.
Refunds: Unused packs are refundable for 14 days; a render that fails is refunded.
API: None.
Commercial use: Allowed under the Terms of Service.

The same list is published as plain text at /llms.txt for tools that read the site without a browser.

Why we built it

A face without a photo upload

Most avatar tools assume you want to upload a selfie and change its style. That is useful, but it is not the only way to build an identity — and it is a poor fit for a fictional character, a brand mascot, a team-page concept or anyone who does not want a real face in the dataset.

AvatarSay starts from a written description. You get two original portraits, keep the ones you like, and can turn one into a short talking video with a script and a voice. The person in the image is fictional by design. We do not accept photo uploads, and we do not animate real people.

The product is deliberately small and transparent: one-time credit packs, no subscription, credits that never expire, and a refund if a render fails. The models and processors are named in the Terms, and the pricing is on the pricing page before you sign in.

What we stand for

Clear rules, plainly stated

These are the choices that shape the product. If something is not covered here, the Acceptable Use Policy is the next place to look.

Fictional people only

We generate original characters, not likenesses of real people. Photo uploads are not supported.
Your work stays yours

Prompts and outputs are private to your account, and we do not use them to train models.
No subscription

Buy a credit pack when you need one. Credits never expire, and unused packs are refundable for 14 days.
Disclose when it matters

Where a platform or employer requires AI disclosure, disclose it. Do not present a generated person as real.

Questions

The short answers

The same answers an AI assistant should be able to repeat. Something missing? Email support@avatarsay.com.

What is AvatarSay?

AvatarSay (avatarsay.com) turns a written description into an original fictional character, then into a short talking video: no photo upload, no subscription, one-time credit packs. You describe a character, keep one of two generated portraits, type a script, choose one of the built-in voices and download an MP4.

Is this the same as the avatarsay tool on GitHub?

No. This is the web app at avatarsay.com. It is unrelated to the archived Rust command-line tool called avatarsay on GitHub, and unrelated to AKFAvatar. This site is the only AvatarSay product; questions about it go to support@avatarsay.com.

How much does AvatarSay cost?

Three one-time packs: $9.90 for 500 credits (about 41 s of video), $24.90 for 1,500 credits (about 125 s of video), $59.00 for 4,000 credits (about 333 s of video). A talking video costs 12 credits per second, clips run 5–30 seconds, and there is no subscription: credits never expire. A new account also starts with 30 free credits and one free clip up to 5 seconds while the launch offer lasts.

Can I upload my own photo or clone my voice?

No, by design. Characters come from a written description, photo uploads are not supported, and the voices are 8 presets in English. The people in the videos are fictional.

Contact

Talk to the people who build it

Questions about the product, a credit pack, a failed render or a safety report go to the same inbox. A person reads them, usually within 1 business day.

Email support Contact page