Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.aytada.app/llms.txt

Use this file to discover all available pages before exploring further.

Ad banners in Aytada are static marketing images built around readable, on-brand text. Because banner performance lives or dies on headline clarity, Aytada uses a typography-aware generation model as the foundation — then composites your actual product into the scene rather than asking the model to hallucinate one. The result is a print-ready banner with your exact product, accurate copy, and a consistent brand feel across every size.

​
What Aytada generates

Each banner request produces a set of images exported at the standard ad sizes:
SizeDimensionsPrimary platforms
Facebook / LinkedIn wide1200 × 628 pxFacebook News Feed, LinkedIn Sponsored Content
Instagram square1080 × 1080 pxInstagram feed, Facebook square placement
LinkedIn standard1200 × 627 pxLinkedIn article header, LinkedIn ads
All sizes are generated from the same campaign context in a single run, so you do not need to manually resize or recompose. Cost: 4 credits per banner.

​
How the pipeline works

The banner pipeline has four stages that run automatically after you click Generate Banner in the Project Studio:
1

Context extraction

Aytada reads your campaign’s brand identity (colors, tone, voice), the ad copy from your script (headline, sub-headline, CTA), and any uploaded product image. If you are generating a standalone banner without a video project, you enter these details directly in the banner module.
2

Typography composition

A typography-aware image generation model renders the background, layout, and all text elements — headline, sub-headline, call-to-action button, and logo placement — as a single cohesive image. This model was chosen specifically because it reasons about text as a design element, producing legible copy at any size rather than distorted or misspelled characters.
3

Product compositing

If you provided a product image, it is cleanly extracted from its background and composited into the generated scene. The compositing model blends shadows, reflections, and perspective so the product looks native to the environment rather than pasted in.
4

Optional 4K upscale

For print or high-DPI digital placements, you can trigger a one-click upscale after generation. The upscaler runs detail reconstruction on the full-resolution image without introducing artifacts or color drift.

​
Tips for best results

Provide a clean product photo with a white, light grey, or neutral background. The compositing step works best when the product edges are sharp and uncluttered. A well-lit studio photo will produce significantly better results than a lifestyle shot with a complex background.
Write your headline and CTA before generating. The AI reads your script’s hook and CTA when composing text. If you customize the copy in the banner module, keep the headline under 8 words and the CTA under 4 words to avoid text overflow at smaller sizes.
The banner generation pipeline uses commercially licensed models for product compositing. Your product images are processed only for the purpose of generating your banners and are not used to train any models.
Typography generation occasionally produces minor spelling variations on unusual product names or URLs. Preview the generated text carefully before downloading for paid placements. You can regenerate at no extra credit cost if the text is incorrect.

​
Using banners in a campaign

Banners are attached to the same project as your video ad. From the Campaign Hub, you can view all assets — video, banners, flyers, and jingles — generated from a single campaign context, ensuring visual and messaging consistency across every channel.

Social flyers

Generate event and promotional flyers using the same product and brand context.

Generate video ads

Create the full cinematic video ad that pairs with your banner campaign.