WebVision Canvas for Phaser

WebVision Canvas lets you pick the brain behind your game code. You choose between a cloud provider, a local model on your PC, or the built-in option. Each choice affects where prompts go.

Choose your model provider

Cloud AI provider

Send your Phaser code to a standard API endpoint on the internet. This option works well when you need strong reasoning for complex refactorings or large scene splits. Your prompts and file text go to that provider. Secrets in tool output are redacted before they reach the model, but you should still avoid putting secrets in files the agent reads.

Local model on your PC

With a local model, prompts and file text stay on the PC.

Built-in model

Use the default model included in the application for quick edits. This is the built-in model, ready to use without your own endpoint or API key, letting you focus on your game design.

Test your connection

Press the Test button to verify your chosen endpoint works correctly. The tool measures response time and checks your context window size.

Pick the right model for the task

Use a bigger model for heavy refactors or complex logic changes. Send hard steps to larger models using the run_on_model tool. Smaller models handle quick edits well.

Quality varies between models, so always review the diff. Use Ask first to pause before accepting changes. This lets you spot errors in your Phaser code before they break your game. You stay in control of every edit.