Overview
Edy brings powerful AI-assisted coding directly into Visual Studio Code through the native Language Model API.
Use Edy to build features, debug code, understand unfamiliar projects, refactor existing code, and work with multiple AI models directly inside VS Code.
Current release: v0.2.5 (Beta)
Features
- Native VS Code Language Model integration
- Access to multiple AI models
- Code generation
- Bug fixing and debugging
- Code explanation
- Refactoring assistance
- Reasoning for complex coding tasks
- Streaming responses
- Secure browser-based sign-in
- Edy credit rates in VS Code's model manager
- Request IDs and usage diagnostics for support
Installation
-
Install the Edy Cloud AI extension.
-
Open the Command Palette:
Ctrl+Shift+P
- Run:
Edy: Sign In
-
Your browser will open. Sign in or create your Edy account.
-
Return to Visual Studio Code.
-
Open GitHub Copilot Chat or the VS Code Chat view.
-
Select an Edy Cloud model.
You're ready to code.
No API key copying or manual provider configuration is required.
Example Prompts
Try asking:
Explain this function.
Refactor this class.
Write unit tests for this file.
Fix this TypeScript error.
Optimize this SQL query.
Convert this JavaScript project to TypeScript.
Developer API
Building your own application with Edy?
Edy also provides developer API keys that let you access supported AI models from your own applications, backends, bots, agents, and developer tools.
Developer API keys are separate from the VS Code sign-in process.
Learn more in the Edy developer documentation:
https://edycode.vercel.app/docs/quickstart
Requirements
- Visual Studio Code 1.125 or later
- An Edy Cloud account
- Internet connection
Privacy
Edy sends prompts to the Edy Cloud API only when you interact with an Edy model.
Authentication credentials are stored securely using Visual Studio Code Secret Storage and are never written to your workspace.
Published builds connect only to the official Edy Cloud service. Workspace settings cannot redirect authentication or API traffic to another server.
Beta Status
Edy is currently in beta.
During this phase, you may see continuous improvements to:
- Model quality
- Reasoning
- Context handling
- Agent capabilities
- Authentication and onboarding
- Reliability
- Performance
Feedback and bug reports are greatly appreciated. Use the troubleshooting
steps below, then contact [email protected] with the Edy request ID.
Troubleshooting
Open View → Output and select Edy Cloud.
The extension logs the HTTP status and Edy request ID for each response without logging your authentication credentials.
When reporting a failed request, include the Edy request ID shown in the output.
The Edy request ID is different from the Client Request Id shown by Visual Studio Code.
Temporary upstream 500, 502, 503, 504, and 529 failures are retried by Edy Cloud up to twice before response output begins.
If capacity remains unavailable, the extension displays a temporary-overload message together with the Edy request ID.