Preparing your workspace...

No files in context
Quick Actions
codmir
Docs
Discord
Get started
API
  • Introduction
  • Installation
  • Your First Project
  • Your First Agent

Control the desktop via X11 tools (xdotool, wmctrl, xdg-open) for URLs, apps, keys, typing, and clicks.

Desktop Automation

Executes desktop actions through X11 tooling.

  • Path: workers/desktop-automation/
  • API (port 8095): /open/url, /launch/app, /window/activate, /key, /type, /click, /intent/run

Examples

  • Open our app: intent text “open our app” → opens http://localhost:3000
  • Open specific URL: POST /open/url {"url":"https://example.com"}
  • Focus window: POST /window/activate {"name_contains":"Firefox"}
  • Type text: POST /type {"text":"hello"}

Environment

  • DISPLAY (X11), DEFAULT_APP_URL, DRY_RUN

Wayland Note

Limited support on pure Wayland; use X11/XWayland for full control.

Open in ChatGPTOpen in ClaudeConnect MCP on CursorConnect MCP on WindsurfEdit on GitHub
codmir
Products
Desktop AppMobile AppIDE ExtensionAI AgentsOverseerVoice ModeCortex ProtocolHive MindGame Dev (Godat)PricingDownload
Solutions
AI Code AgentsError TrackingSelf-HealingDeploy TrackingAI Tracking SDKWorkflow AutomationTesting FrameworkTime TrackingVoice Control
Platform
ProjectsTickets & BoardsAgent MarketplaceIntegrationsTemplatesApp HubAdmin PanelAnalytics
Integrations
GitHubSlackDiscordVercelSentryRailwayLinearAll Integrations
Developers
DocumentationAPI ReferenceSDKCLIDeveloper PortalMarketplaceMCP ConnectorsWebhooksExamplesnpm Packages
Resources
BlogChangelogCommunityGet StartedSupportStatusFeedback
Company
AboutCareersContact SalesKickstarterReferral Program
Legal
Terms of UsePrivacy PolicyEULATrademarkSecurity
GET IT ONTestFlightGet it on Google Play

Codmir™ 2026. All rights reserved.

Free & open source SDK — Apache 2.0 License