MAGIC WINDOW V1.01
Capture. Paint. Direct. Generate.
AI-assisted look development directly from the Unreal Editor viewport
Magic Window is an Unreal Editor plugin for viewport-driven AI look development.
Capture the active Unreal viewport, mark it up directly inside Unreal, add quick text or shape notes, define cinematic direction, and generate screenshot-guided image ideas without leaving the editor.
The plugin is built for artists and creators who want to explore lighting, mood, composition, art direction, material polish, and cinematic shot ideas faster while still using Unreal as the controllable visual base.
Magic Window supports external AI image generation through user-provided API keys. This build supports Google Gemini and OpenAI image generation via API. The plugin does not include AI credits; generation costs are billed by the selected provider.
What Magic Window Does
Magic Window turns your Unreal viewport into a directed AI lookdev workspace:
Core Features
Active Unreal Editor viewport capture
Screenshot-guided AI image generation
Google Gemini and OpenAI API support
User-provided API key workflow
Capture, Iteration, and Reference creative edit targets
Reference Tray for captures, generated results, and external references
Draw-on-capture, draw-on-result, and selected-reference workflows
Brush, eraser, text, triangle stamp, and square stamp tools
Keep, Remove, Select, and Reset mark modes for guided edits
Dedicated Compare Window with independent source / result display
Vertical source / result comparison with equal-size frames
Proportional fit rendering so full images remain visible
Viewpoint, lens / camera, mood, and time / light preset helpers
Social-format and cinematic aspect presets
2K and 4K output selectors where supported by the selected model
PNG and JPG output options
Startup AccessPoint connection check
Session cost estimate panel
Project-local output folders
Updated README and user guide for the V1.01 workflow
The goal is not to replace Unreal rendering. The goal is to speed up visual exploration before committing to a heavier production path.
What’s New in V1.01
Magic Window V1.01 focuses on a cleaner creative workflow, stronger image-target handling, a more useful Reference Tray, and better comparison behavior.
Added explicit Creative Edit target handling for Capture, Iteration, and Reference images
Added Reference Tray workflow for capture, result, and external image references
Added selected-reference editing support for guided masks and selection-style reference intent
Added dedicated Compare Window context so viewport captures and generated iterations remain separate
Updated Compare Window to use vertical source / result comparison with equal-size frames
Improved proportional fit rendering so the full image remains visible in Compare Window
Added startup AccessPoint connection check so invalid keys are not shown as configured
AI Provider / Billing Notes
Current production integrations in this build:
Google Gemini
OpenAI image generation via API
Provider and model behavior can vary. Some integrations may feel more interpretive with rough concept sketches and blockout-style guidance, while others may respond more conservatively and shine when refining an already established scene or hero asset.
Magic Window is designed to support both workflows, but users should expect different visual behavior depending on the selected provider and model.
The plugin includes session-aware preview panels, simple cost estimation, output-folder access, social-format size presets, 2K and 4K resolution options where supported by the selected model, and a clean editor-native UI designed for repeat use.
Magic Window does not include AI generation credits.
Users must provide their own API key and an active provider account with billing or credits enabled. API usage costs are billed by the selected provider, not by the plugin.
The built-in cost estimate is informational and intended for session planning. It is not a provider invoice.
Important OpenAI note:
ChatGPT subscriptions do not include OpenAI API usage
OpenAI API billing must be activated separately
Access to GPT image generation may require provider-side organization verification
Important Google note:
Google Gemini API usage is billed through the user’s Google AI account
Provider pricing, rate limits, and preview model availability can change over time