Connect your AI tool
Install paths for every supported AI client. Pick yours and follow the steps.
All AI clients connect to the same MCP server: https://mcp.shiftcontrol.io/mcp. OAuth discovery is at https://mcp.shiftcontrol.io/.well-known/oauth-authorization-server, and Dynamic Client Registration is supported. You'll be prompted to sign in to ShiftControl in your browser the first time the AI calls a tool.
Claude Desktop (macOS, Windows, Linux)
Open your Claude Desktop config
| Platform | Path |
|---|---|
| macOS | ~/Library/Application Support/Claude/claude_desktop_config.json |
| Windows | %APPDATA%\Claude\claude_desktop_config.json |
| Linux | ~/.config/Claude/claude_desktop_config.json |
Add ShiftControl to the mcpServers block
{
"mcpServers": {
"shiftcontrol": {
"command": "npx",
"args": ["mcp-remote", "https://mcp.shiftcontrol.io/mcp"]
}
}
}
Restart Claude Desktop
The first tool call opens ShiftControl in your browser to sign in. Approve, and the tool call completes.
Claude AI (Chat, CoWork, Code)
- Open Claude → Customize → Connectors
- Click the + button and select Add custom connector
- Name:
shiftcontrol. Remote MCP Server URL:https://mcp.shiftcontrol.io/mcp - Click Add
Claude Code
claude mcp add shiftcontrol --transport http https://mcp.shiftcontrol.io/mcp
Cursor
- Settings → MCP → Add new MCP Server
- Name:
shiftcontrol. URL:https://mcp.shiftcontrol.io/mcp - Save and restart Cursor
Windsurf
- Settings → MCP Servers → Add server
- URL:
https://mcp.shiftcontrol.io/mcp - Restart Windsurf
Cline (VS Code extension)
- Settings → MCP Servers → Add server
- Name:
shiftcontrol. URL:https://mcp.shiftcontrol.io/mcp
Continue.dev
Edit ~/.continue/config.json:
{
"experimental": {
"modelContextProtocolServers": [
{
"transport": { "type": "http", "url": "https://mcp.shiftcontrol.io/mcp" },
"name": "shiftcontrol"
}
]
}
}
ChatGPT
ChatGPT's native MCP support is rolling out. Until it lands:
- Custom GPTs — use the ShiftControl OpenAPI as an Action:
https://api.shiftcontrol.io/api-docs/openapi.json - ChatGPT Desktop — check OpenAI's release notes for current MCP status
Gemini CLI
gemini mcp add shiftcontrol --url https://mcp.shiftcontrol.io/mcp
Generic OAuth-MCP client
If your client implements the Model Context Protocol with OAuth, point it at:
| Setting | Value |
|---|---|
| Endpoint | https://mcp.shiftcontrol.io/mcp |
| OAuth discovery | https://mcp.shiftcontrol.io/.well-known/oauth-authorization-server |
| Dynamic Client Registration | Supported |
Verify the connection
After install, ask the AI to run list_apps. You should see your ShiftControl apps come back. If you don't, the troubleshooting section in the skills repo INSTALL.md is the authoritative source.
What's next
- Skills — install pre-built workflows on top of the MCP server