Recraft

Prompt

PROJECT: AI-NATIVE VISUAL DESIGN AND BRAND CREATION STUDIO

Objective

Build a sophisticated AI-powered creative design platform that transforms text instructions, reference images and visual concepts into production-ready artwork.

The application should combine generative artificial intelligence with professional editing tools, allowing creators to generate, modify, refine, organize and export graphics without switching between multiple applications.

The platform must feel like an intelligent creative environment rather than a conventional AI chatbot or a basic text-to-image generator.

1. Premium Design Workspace

Create an infinite visual canvas where users can arrange generated images, vector graphics, brand assets, mockups and creative references.

Implement smooth zooming, panning, object selection, drag-and-drop positioning, snapping guides, alignment controls, layers and frame management.

Build a contextual editing interface consisting of a central canvas, collapsible creation panel, contextual properties panel and minimal floating toolbar.

Provide light and dark interface modes while prioritizing professional usability.

2. Advanced AI Image Generator

Integrate high-quality image generation through authenticated AI model APIs.

Allow users to generate images using natural-language prompts describing subjects, lighting, camera angles, composition, mood, materials, visual styles and color palettes.

Support configurable aspect ratios, output dimensions, image quantities and model selection based on available provider capabilities.

Include image previews, prompt history, variations and generation progress.

Allow users to generate several creative interpretations from the same prompt and compare them before selecting a final direction.

3. AI Vector Graphics Generator

Develop a dedicated system for creating scalable vector illustrations, logos and icons.

Use a model capable of producing structured vector outputs, or implement a validated raster-to-vector conversion pipeline when native vector generation is unavailable.

Allow users to inspect and edit generated SVG paths, fills, strokes, curves and colors.

Support resizing without loss of quality, SVG export, element grouping and transparent backgrounds.

The vector editor must maintain meaningful editing capabilities rather than simply embedding raster images inside SVG containers.

4. Intelligent Brand Identity Generator

Build an AI brand design workspace capable of developing coordinated visual assets.

Users should provide a brand name, industry, target audience, personality, preferred colors and creative direction.

Generate a proposed brand identity containing:

- Logo concepts and alternative arrangements.
- Brand color palette with hexadecimal values.
- Suggested typography combinations.
- Icon and illustration style.
- Social media visual directions.
- Marketing graphics and product imagery.
- Basic brand guidelines.

Allow users to approve, reject or regenerate individual assets without losing the broader creative direction.

Create a persistent brand kit that subsequent generations can reference to maintain visual consistency.

5. AI Image Editing System

Implement professional editing capabilities powered by appropriate image-processing and generative models.

Support background removal, object removal, masked inpainting, outpainting, background replacement, creative restyling and intelligent resizing.

Allow users to select a specific image region and describe the desired modification using natural language.

Preserve unaffected regions wherever the model and editing pipeline support precise localized changes.

Display before-and-after comparisons and provide reliable undo and redo functionality.

6. Design Style Intelligence

Create a reusable visual style system.

Users should be able to upload reference images or select visual presets to establish design characteristics such as composition, texture, palette, lighting and illustration treatment.

Represent these preferences as reusable style configurations.
Apply saved styles to future generations while allowing creative variations.

Clearly distinguish style references from exact reproduction, and implement safeguards against unauthorized replication of protected designs.

7. AI Product Mockup Studio

Develop a mockup generation environment that turns flat artwork into realistic product presentations.

Support mockups for packaging, clothing, bottles, books, digital devices, stationery and marketing materials.

Allow users to upload artwork, select a product category and generate a realistic presentation scene.

Include controls for perspective, positioning, shadows, materials, environmental lighting and visual context where supported.

Preserve the original artwork's important text and proportions.

Provide a preview-and-approval workflow before exporting.

8. Conversational AI Design Agent

Integrate an intelligent assistant using Claude API or another suitable reasoning model.

The design agent should understand the current project, selected canvas elements, saved brand identity and previous user instructions.

Allow commands such as:

"Create five logo concepts for a premium travel company."

"Make this illustration more editorial while preserving its colors."

"Turn these assets into a consistent marketing campaign."

"Adapt this visual identity for a mobile application."

The agent should convert natural-language requests into structured editing operations and supported AI-generation calls.

Every action must be observable, reversible and subject to appropriate user permissions.

9. Design Comparison and Refinement

Create a workspace for comparing multiple design outputs.

Support side-by-side comparisons of two or more generated images, including model differences, prompt changes and style variations.

Allow users to save preferred results, refine a selected design and maintain a complete revision history.

Include optional AI-assisted quality checks for layout consistency, contrast, readable text and visual alignment.

Quality assessments should be presented as suggestions, not guaranteed objective judgments.

10. AI-Assisted Typography and Layout

Provide tools for creating posters, social graphics, marketing campaigns and other compositions containing both images and text.

Develop an editable text layer system supporting fonts, sizes, spacing, line height, alignment, hierarchy and positioning.

Let users request layout improvements through natural language.

Avoid relying on generated image text when precise editable typography is required.

Implement real text layers for accurate corrections and professional export.

11. Creative Project Management

Allow users to organize work into projects, collections and brand workspaces.

Implement asset search, favorites, generation history, tagging, duplication and reusable templates.

Provide secure sharing links and optional collaboration permissions.

Store original files, generated variations, prompts and configuration metadata.

12. AI Model Architecture

Build a modular backend that supports multiple image generation and editing providers.

Introduce an adapter layer to normalize requests, model parameters, generation status, file retrieval and errors.

Keep all provider API keys on the server.

Track generation costs and usage accurately.

Only expose settings and output formats that are actually supported by the selected model.

13. Technical Stack

Use Next.js, React, TypeScript, Tailwind CSS and Framer Motion for the application interface.

Implement a performant canvas using an appropriate graphics rendering library.

Use PostgreSQL with Prisma for project metadata, object storage for generated assets and Redis-backed queues for longer-running generation operations.

Integrate AI image and vector APIs, a language-model API for the design assistant and appropriate image-processing libraries.

Provide authenticated user accounts, secure asset access, rate limits, error handling and deployment documentation.

14. Visual Experience

Design an original, elegant application with a strong editorial identity.

Use a restrained neutral palette, sharp typography, carefully positioned controls and subtle interaction feedback.

Let generated artwork provide the visual richness of the interface.

Avoid generic SaaS dashboards, unnecessary gradients, oversized decorative cards and visual effects that compete with the canvas.

Build a responsive experience for desktop and tablet users.

15. Final Deliverables

Produce a complete application containing a premium landing page, authenticated design studio, infinite canvas, AI image generator, editable vector workflow, reusable brand kit, image editor, mockup studio, conversational design assistant and asset export system.

Implement real AI integrations rather than simulated generation responses.

Provide a clean repository structure, environment-variable documentation, setup instructions, deployment configuration and useful error states.

Final Vision

Create an AI-native graphic design studio where users move seamlessly from inspiration to finished creative assets.

The core principle is simple:

AI generates the possibilities. Designers control the final result.