built-in browser with device emulation and AI capture
Canopy includes a Chromium-based browser tab alongside the terminal. Switch between device viewport presets, save credentials in the OS keychain, and capture elements or screen regions to send directly to a running Claude Code session.
element and screenshot capture for AI agents
Click the element picker to hover over a page element and capture its HTML, or draw a region to take a screenshot. Canopy sends the captured data to the active Claude Code or Gemini CLI session in the same worktree.
- Element picker captures the HTML of the selected element
- Region capture saves a PNG and sends the file path to the agent
- Session picker when multiple agents are running
device emulation presets
Switch between viewport presets using Chrome DevTools Protocol. The viewport changes instantly without reloading the page. Custom viewports are saved across sessions.
- iPhone SE, iPhone 14 Pro, iPad Mini, iPad Pro 11", Pixel 8, Samsung Galaxy S24
- Device pixel ratio override and mobile flag
- Custom viewport sizes saved in preferences
credential storage and favorites
Saved credentials are encrypted via the OS keychain (Keychain on macOS, Credential Manager on Windows). An autofill icon appears in password fields for one-click login. Favorites let you bookmark URLs for quick access.
- Credentials encrypted via Electron safeStorage and the OS keychain
- Autofill icon injected into password fields
- Favorites list for bookmarked URLs
questions about the built-in browser
- What rendering engine does the browser use?
- The built-in browser uses Chromium via Electron's webview tag. It behaves like Chrome for web compatibility.
- Is browser state isolated per worktree?
- Cookies and local storage are shared across all browser tabs in Canopy. Each worktree opens its own browser tab, but login sessions carry over between them.