One photo in · Reads at 28 pixels · Nothing to install
Bold outlines, flat colour, one big expression — a chat icon that still reads when the platform shrinks it. Free to start, no vector software.
A real run through the model this page uses, on a square source photo. Drag the divider: the curls, the beard and the hoodie survive the redraw, but the line weight and the expression are rebuilt for the small size. Every output is AI-generated, so your result varies with your photo — treat this as a guide to what the tool does, not a fixed result you will get back exactly.


Drag to compare before and after
An emote is a tiny reaction icon that lives in a chat line on Twitch, Discord, YouTube or Slack. It is not a small illustration — it is a drawing designed backwards from the size it will be seen at.
Three constraints do all the work. The outline has to be thick and unbroken, because at icon size a thin line disappears into the chat background and the shape falls apart. The palette has to stay small and flat, since gradients turn to mud once the platform resamples the file down. And the expression has to be pushed much further than feels natural at full size — squeezed-shut eyes, a mouth open to the corners, brows all the way up.
A subtle face is a grey blob in chat, which is why a photo pasted into an emote slot never works.
The redraw runs on the Nano Banana image model in the browser, with free credits to start. Drop in the photo, name the reaction, and the credit estimate settles before you commit; a failed run is refunded automatically unless it broke the content policy. Feed it a square head-and-shoulders shot with your face large in the frame and one clear expression — the more of the frame your face occupies, the more detail survives the shrink.
Full-body shots, busy backgrounds and neutral expressions all give the redraw less to work with. Platforms want transparent or flat-background PNGs at 112, 56 and 28 pixels, so generate large and resize down rather than the other way round.
Creative engine
Upload a square photo with one clear expression, say which reaction it should read as, and the credit estimate updates before you generate.
Thick outlines, flat cel colour and an exaggerated expression — drawn for the 28-pixel version, not shrunk into it.
Runs in the browser with free credits — no vector software, no commission queue, no account wall before you see a result.
Neighbouring pages cover the rest of a channel's art: a die-cut look in the AI sticker generator, a captioned reaction in the speech bubble meme generator, and a collectible card layout in the Pokemon trainer card maker.
From one photo to a chat icon, with the cost shown before you commit.
Head and shoulders, face filling most of the frame, one unmistakable expression. Everything past your collar is going to be cropped away anyway.
Crying with laughter, deadpan stare, rage, shock, smug. Say the reaction rather than the style and the drawing follows it.
Generate, then resize to 112, 56 and 28 pixels before uploading. A failed run is refunded unless it broke the content policy, so a second reaction costs you nothing.
Streamers filling a subscriber tier, servers building an in-joke, teams that live in chat.
Unlocking emote slots and then having nothing to put in them is the standard bottleneck. A matching set beats one commissioned piece and five empty slots.
The face somebody pulled on voice chat, immortalised where the joke actually happens — and a matching sticker for the same server.
A tiny cartoon of each person makes standup threads readable at a glance, which is a real thing that a decorative avatar is not.
Common questions about turning a photo into a chat icon with Kavel.
Twitch takes 112, 56 and 28 pixel PNGs; Discord takes 128; YouTube takes 32. Generate at the largest size the tool offers and scale down from there — a small file scaled up is soft, and soft is exactly what kills an icon at chat size.
The generator returns a flat background you can knock out afterwards; ask for a plain solid colour behind the bust and it separates cleanly. Platforms accept flat backgrounds too, though transparency is what lets a set sit properly on both light and dark themes.
Yes, and that is the point of running it more than once. Use the same source photo, name a different reaction each time and keep the same style words in the prompt — that is what makes a row of them look like one set instead of five commissions.
Almost always because the drawing was too detailed for the size. Thin lines, many colours and small features vanish under resampling. Ask for heavier outlines, fewer colours and a bigger expression, and crop tighter on the face before you run it again.
Yes. Pets work particularly well because their features are already exaggerated, and the same redraw handles an object or a mascot. What it needs from any source is a clear subject that fills the frame and one readable expression.
The estimate is on screen before you press generate and moves with the format and resolution you pick, so nothing runs unpriced. Failed runs are refunded unless they broke the content policy, and a new account starts with enough free balance to fill a couple of slots first.
Read the terms for the current position on outputs before building a subscriber tier on them. Worth knowing separately: platforms have their own review process and reject uploads that carry someone else's trademark or a real brand logo, whatever the terms of the tool that drew it.
It should — those are the features regulars use to recognise you, so the redraw carries hair, facial hair, glasses and skin tone across. Small accessories can drop out at this size on purpose; if one matters, say it in the prompt and it comes back.
One credit pool covers Nano Banana images and Seedance and Veo video. The cost shows before every run, and failed jobs are not charged. Use a subscription for ongoing work, or a one-time pack when you just need to top up.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
One-time top-ups — buy extra credits any time you run low.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
Charges appear as “KAVEL AI” on your card statement. You can cancel any time from Settings → Billing; cancellation takes effect at the end of the period you already paid for.
Powered by
Use one balance across every supported image and video model — ten of them today, listed below — and check the credit cost before each request.
Video models
Image models
Model credit guide
You're only charged for successful generations. The exact estimate in the generator varies by model, length, resolution, audio, and number of images.
| Type | Model | Credit cost |
|---|---|---|
| Video | Seedance 2.0 | 5s 720p image-to-video ≈ 188 credits; text-to-video ≈ 308 credits. Scales with resolution and length. |
| Video | Seedance 2 Fast | Faster and lower cost. 5s 720p text-to-video ≈ 248 credits. |
| Video | Seedance 2 Mini | The cheapest Seedance tier. 5s 720p ≈ 154 credits, 5s 480p ≈ 72 credits. |
| Video | Seedance 1.5 Pro | Audio doubles the rate. 5s 720p ≈ 27 credits silent, ≈ 53 with audio; 1080p ≈ 57 and ≈ 113. |
| Video | Veo 3.1 | Billed per video, not per second (Lite tier). About 45 credits at 720p and 53 at 1080p. |
| Video | Kling 3.0 | Audio raises the rate. 5s 720p ≈ 105 credits silent, ≈ 150 with audio; 1080p ≈ 135 and ≈ 203. |
| Video | MiniMax H3 | Fixed 2K, no resolution ladder. Priced per second — a 5s clip ≈ 158 credits. |
| Image | Nano Banana 2 | Generate or edit from text and images. About 8 credits per 1K image, 12 at 2K, 18 at 4K. |
| Image | Nano Banana Pro | Consistent run times across generations. About 12 credits per 1K or 2K image, 21 at 4K. |
| Image | Nano Banana 2 Lite | Faster, simpler variant. A flat 8 credits per image at every resolution. Start here to test. |
| Image | GPT Image 2 | The lowest-cost image model here. About 3 credits per 1K image, 6 at 2K, 12 at 4K. |
| Image | Seedream 5.0 Lite | Flat pricing — about 8 credits per image whether you generate or edit. |
One photo in, a bold cartoon icon out — readable at the size chat actually shows it. Nothing to install and no commission queue to join.