⌨ Keyboard shortcuts available
G — waiting for next key…
AI Creative Studio · Powered by Gemini
PRISM

Generate
ad creatives
in seconds.

Describe your idea. Prism uses Google Gemini to generate the background, write the copy, and place the text — giving you a pixel-perfect creative you can customize and export.

3 Canvas formats
2 Creative modes
0 Data stored
Iterations
pradeepbhandari.com/prism/app
CANVAS Instagram Post Instagram Story Landscape Ad MODE Layered Composer AI-Integrated LAYERS T Headline T Subheadline T CTA Button Export PNG EXPLORE MORE Discover new horizons this season Book Now → Describe your creative idea… AI → STYLE Font Family Inter Size 72px Color Align Shadow Opacity

What Prism does

Everything you need to
ship creatives faster

Six tightly integrated capabilities — from AI ideation to pixel-perfect export — in a single browser tab.

AI Layout Generation

Describe your campaign idea in plain language. Gemini 2.0 Flash analyzes your prompt and produces a structured layout: background scene concept, headline, subheadline, and CTA copy — all tailored to your chosen canvas format.

Photorealistic Background Generation

Gemini's image model renders a clean, text-free background at the correct aspect ratio. Whether you need a mountain vista, a sleek product shot, or an abstract gradient — you get a production-ready image, not a stock photo placeholder.

Drag-and-Drop Text Composer

Every text element lives on its own Konva.js canvas layer. Drag to reposition, use the transform handles to resize, and dial in font family, size, weight, color, alignment, shadow, and opacity — without touching any code.

AI-Integrated Poster Mode

Prefer typography that feels like part of the art? Switch to AI-Integrated mode: the model generates a full poster with text naturally embedded in the scene. Gemini Vision then detects every text zone so you can fine-tune placement and style as editable layers on top.

Inpainting & Text Removal

When the AI bakes typography into an image but you want a cleaner base, Prism can re-generate the image instructing the model to remove the detected text — producing a clean background canvas ready for your own layered text.

Client-Side PNG Export

Export merges the background image and all text layers into a single flat PNG directly in the browser — no round-trip to the server, no file size limits, no data retained. Click export, get your file.

Simple by design

From blank to export
in three steps

01

Choose your format

Select the canvas that matches your placement: Instagram Post (1080×1080), Instagram Story (1080×1920), or Landscape Ad (1920×1080). Prism applies the correct aspect ratio to every generation call automatically.

02

Describe your creative

Type a prompt like "luxury watch brand, midnight blue background, stars, premium feel". Prism asks Gemini to write the copy and the background scene description, then renders the image — all in one click.

03

Customize and export

Drag text layers to the perfect position. Adjust the font, size, color, and weight. When you're happy, hit Export PNG — the final file downloads directly to your device at full resolution.

Your Prompt
Gemini Layout
Background Image
Text Layers
PNG Export

Two creative approaches

Pick the mode that fits
your workflow

Both modes use the same Gemini backbone. The difference is how much control you want at the typography stage.

Mode A

Layered Composer

The background is generated text-free. Your copy is placed as independent, fully editable canvas layers that you control completely.

  • Clean, text-free AI background at full resolution
  • AI-suggested copy pre-loaded as draggable layers
  • Pixel-perfect control over every character's position and style
  • Ideal when your brand has strict typography guidelines
  • Works best with photography-style and abstract backgrounds
BEST FOR

Brand-consistent ads, product launches, campaign creatives where typography needs to match existing brand identity.

Mode B

AI-Integrated

Gemini generates the entire poster — background and typography — as a single cohesive image. Vision AI then extracts the text zones so you can still tweak them.

  • Full poster rendered in a single generation pass
  • Typography is baked into the image as artistic design
  • Gemini Vision detects and maps every text element automatically
  • Detected zones become editable overlay layers for fine-tuning
  • Inpainting removes baked text if you prefer clean separation
BEST FOR

Posters, editorial content, artistic social creatives where the typographic treatment is itself a visual statement.

Platform-ready dimensions

Every format,
perfectly sized

Instagram Post

1080 × 1080 px

Square format — the gold standard for feed posts. High visual impact, equal weight in both portrait and landscape feeds.

Feed posts Carousels Reels covers

Instagram Story

1080 × 1920 px

Full-screen vertical format that takes over the entire display. Maximum immersion, perfect for limited-time offers and announcements.

Stories Reels TikTok ads

Landscape Ad

1920 × 1080 px

Cinema-wide format built for YouTube pre-rolls, LinkedIn feed ads, Google Display, and desktop banner placements.

YouTube ads LinkedIn Display ads

Under the hood

Built on
best-in-class AI

Prism uses Google Gemini models at every stage of the creative pipeline — choosing the right model for each task to balance speed, quality, and cost.

gemini-2.0-flash

Layout planning, copy generation, text zone detection via Vision. Fast, cheap, and razor-sharp at structured reasoning.

gemini-3.1-flash-image-preview

Background and full-poster image generation. Native multi-modal output delivers high-fidelity images from text prompts.

Konva.js · Intervention Image

HTML5 canvas layer compositing for text, transforms, and export. PHP-side image resizing ensures pixel-perfect dimensions.

Your API Key, Your Data

Prism uses your own Google Gemini API key — stored for the session only and never persisted to a database. All generation calls go directly from the server to Gemini using your key.

No images stored on this server
No account or sign-up required
API key cleared when session ends
Export happens entirely client-side
Bring your own Gemini API key

HOW TO GET A GEMINI KEY

Create a free API key at aistudio.google.com. The free tier includes generous image generation quota — enough for dozens of creative sessions per day.

Common questions

FAQ

Yes. Prism is free to use — you only need a Google Gemini API key. Google provides a generous free tier through AI Studio. You are billed by Google (not by this site) only if you exceed the free quota.

No installation. Prism runs entirely in the browser. Open the Studio link, paste your API key once per session, and you are ready to generate.

Nothing is stored. Generated images exist only in your browser session. The export writes the PNG directly to your device — nothing is saved server-side.

Gemini's image generation model is computationally intensive. Generation time varies with server load. The preview animation will spin until the image arrives.

Usage rights for AI-generated content depend on Google's Gemini usage policies. Check the Gemini API Terms of Service for the latest guidance on commercial use.

Prism includes a curated set of system and Google Fonts: Inter, Playfair Display, Montserrat, Raleway, Oswald, Merriweather, and more — covering serif, sans-serif, and display styles.

PRISM

Ready to create your
first AI creative?

Grab a free Gemini API key, open Prism Studio, and go from prompt to export in under a minute.

Open Prism Studio See more projects

FREE TO USE · BRING YOUR OWN API KEY · NO ACCOUNT REQUIRED