00widget MCP server
{
"mcpServers": {
"00widget": {
"type": "http",
"url": "https://api.00widget.com/mcp"
}
}
}{
"mcpServers": {
"00widget": {
"type": "http",
"url": "https://api.00widget.com/mcp"
}
}
}{
"mcpServers": {
"00widget": {
"type": "http",
"url": "https://api.00widget.com/mcp"
}
}
}{
"servers": {
"00widget": {
"type": "http",
"url": "https://api.00widget.com/mcp"
}
}
}{
"mcpServers": {
"00widget": {
"type": "http",
"url": "https://api.00widget.com/mcp"
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsonversion 1.0.0
What it does
00Widget exposes tools to publish status cards and Live Activities to iOS widgets, the Lock Screen, and the Dynamic Island. It's designed for Claude, ChatGPT, automation scripts, and coding agents that need a native Apple output channel without building UI code.
- Publish build status, job progress, or queue depth from a CI pipeline to an iOS widget that updates automatically.
- Show real-time state (solar production, delivery tracking, charge cycles) on the Lock Screen via Live Activities with clear end times.
- Let agents like Claude Code or home-automation scripts update dashboard cards by posting structured state to a REST API, with no SDK or UI code needed.
Tools this server exposes
| Tool | Does |
|---|---|
| publish_card | POST a dashboard card with one of five templates (summary, progress, list, action, chart) to display state on iOS widgets; requires stable id and avoids secrets |
| publish_live_activity | POST a Live Activity session to the Lock Screen or Dynamic Island for time-bounded operations with a clear end time. |
| batch_upsert_cards | POST multiple related cards in one call to efficiently update related snapshots across a dashboard. |
| run_action | POST to trigger a backend-defined action via an ActionDefinition, such as a button press from a widget. |
| get_cards | Retrieve all published dashboard cards for the authenticated tenant. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.