image-extender  by boona13

AI image extension and 2D game art generation

Created 2 weeks ago

New!

978 stars

Top 37.5% on SourcePulse

GitHubView on GitHub
Project Summary

Summary

This open-source web application addresses AI-powered image outpainting and 2D game asset generation. It enables users to seamlessly extend images in any direction and create diverse game art assets like parallax backgrounds, autotiles, sprite animations, and props, benefiting game developers and artists seeking efficient asset creation pipelines.

How It Works

The core functionality leverages Google's Gemini image models accessed via OpenRouter, with user API keys stored client-side. It employs a Poisson-blending pipeline for mathematically invisible seams between original and AI-generated content, including pre-correction for color drift. The application offers five specialized studios: Extender for general outpainting with a best-of-3 variant picker, Parallax Studio for multi-layer backgrounds, Tile Studio for autotiles, Sprite Studio for character animations, and Props Studio for decorative elements, each with tailored AI workflows and compositing techniques.

Quick Start & Requirements

Installation involves cloning the repository, running npm install, and then npm run dev. The application is accessible at http://localhost:3000. A primary requirement is an OpenRouter API key, which is stored locally in the browser's localStorage. Costs are approximately $0.03 per Gemini extension. Further details and API key acquisition can be found at openrouter.ai/keys.

Highlighted Details

  • Five Specialized Studios: Offers distinct workflows for Extender (outpainting), Parallax Studio (multi-layer backgrounds), Tile Studio (13-tile autotiles), Sprite Studio (character animations), and Props Studio (decoration sprites).
  • Advanced Seam Blending: Utilizes Poisson blending with a grown mask and Gauss-Seidel iterations for seamless integration, coupled with color drift pre-correction.
  • AI-Driven Game Art Pipelines: Features include role-aware prompts for parallax layers, deterministic corner reconciliation for tiles, a two-pass anchor/sheet workflow for sprite identity preservation, and an "art director -> painter" pipeline for props to ensure variety.
  • Client-Side API Key Management: User API keys are stored exclusively in the browser's localStorage, enhancing privacy as they are never sent to or stored on the server.
  • Keyboard-Centric Interface: Designed for efficient workflow with extensive keyboard shortcuts for actions like extending, cycling variants, and accepting/discarding generations.

Maintenance & Community

The provided README does not contain specific details regarding notable contributors, sponsorships, community channels (e.g., Discord, Slack), or roadmap links.

Licensing & Compatibility

The project is released under the MIT License. This license generally permits commercial use and integration into closed-source projects without significant restrictions, provided the license terms are followed.

Limitations & Caveats

The application's functionality is dependent on the availability and cost of the OpenRouter API and the chosen Gemini models. While client-side API key storage enhances privacy, it requires users to manage their own keys and associated costs. The quality and performance of generated assets are directly tied to the capabilities of the underlying AI models.

Health Check
Last Commit

1 week ago

Responsiveness

Inactive

Pull Requests (30d)
3
Issues (30d)
0
Star History
982 stars in the last 17 days

Explore Similar Projects

Starred by Peter Norvig Peter Norvig(Author of "Artificial Intelligence: A Modern Approach"; Research Director at Google).

NanoBananaEditor by markfulton

0.3%
686
Advanced AI image generation and editing platform
Created 9 months ago
Updated 4 weeks ago
Feedback? Help us improve.