Guest bindings for tauri-plugin-mcp — lets AI agents interact with and debug Tauri apps
Tauri MCP server exists for a simple reason — assistants are far more useful when they can act on Tauri directly instead of describing what you should do. Guest bindings for tauri-plugin-mcp — lets AI agents interact with and debug Tauri apps.
A Tauri plugin and MCP server that allow AI agents such as Cursor and Claude Code to interact with and debug your Tauri application through screenshots, DOM access, input simulation, and more.
Once Tauri is connected, these are the calls the assistant has available:
take_screenshot — Captures a screenshot of an application window. Saves full image to disk with small thumbnail inline (optimized for token efficiency)query_page — Inspects the current page. Modes: map (structured element refs), html (raw DOM), state (URL/title/scroll/viewport), find_element (CSS pixelclick — Clicks via selector (ref, id, class, css, tag, text) or at raw x/y coordinates. Selector-based left clicks dispatch synthetic pointer events at theread_text — Reads visible text of elements matching a CSS selector: per-element {tag, text, visible, attrs?}, whitespace-collapsed, output hard-capped with ainspect_element — Bounding rect, computed styles (default layout/color set or explicit style_props), classList and attributes — for visual/layout QA without execute_jsdispatch_pointer — Synthetic pointer/mouse gesture chains (click, dblclick, down, up, hover, drag) on the exact matched element, no retargeting. Element-relative offsetapp_bridge — Calls helpers the host app registered via window.MCP_BRIDGE.register(name, fn, description) — list to discover, call to invoke (async supportedtype_text — Types text into the page. Supports a fields array for bulk form fill, selector targeting, or typing into the focused element. Selects options inpress_key — Presses keyboard keys the typing tools can't express: Escape, Enter, Tab, arrows, Backspace/Delete, F-keys, and modifier chords (cmd+a). Emulatesmouse_action — Non-click mouse actions: hover, scroll (by direction/amount/to element/to top/bottom), drag (start to end coordinates)navigate — Webview navigation: goto (URL), back/forward (with optional delta), reloadexecute_js — Runs arbitrary JavaScript in the webview. Returns the last expression's value; promises are awaited. The universal escape hatchtauri-plugin-mcp-server on npm is all you need. Most clients run it directly, so configuration is a few lines and a restart.
You will need 5 environment variables: TAURI_MCP_IPC_PATH, TAURI_MCP_CONNECTION_TYPE, TAURI_MCP_TCP_HOST, TAURI_MCP_TCP_PORT, TAURI_MCP_AUTH_TOKEN. The server will not start without them, which is usually why the tools fail to appear on a first run. Keep credentials in your client's env block or a secrets manager rather than in a file you might commit.
Plenty of browser automation servers cover similar ground. The differences that matter in practice are scope of access and how much setup stands between you and a working tool call. Tauri's toolset — take_screenshot, query_page, click and 11 more — is a fair guide to whether it matches your workflow. It is maintained by GitHub Actions; worth a glance at recent repository activity before you build anything load-bearing on it.
We check each listing at SyncDev against the project's documentation before it goes live — if something here drifts out of date, it is a bug worth reporting.
| Tool | What it does |
|---|---|
| take_screenshot | Captures a screenshot of an application window. Saves full image to disk with small thumbnail inline (optimized for token efficiency). |
| query_page | Inspects the current page. Modes: map (structured element refs), html (raw DOM), state (URL/title/scroll/viewport), find_element (CSS pixel coordinates for clicking), app_info (app metadata, windows, monitors). |
| click | Clicks via selector (ref, id, class, css, tag, text) or at raw x/y coordinates. Selector-based left clicks dispatch synthetic pointer events at the element (no OS permissions needed); right/middle/double and raw x/y clic |
| read_text | Reads visible text of elements matching a CSS selector: per-element {tag, text, visible, attrs?}, whitespace-collapsed, output hard-capped with a truncated flag. The structured replacement for execute_js text-scraping sn |
| inspect_element | Bounding rect, computed styles (default layout/color set or explicit style_props), classList and attributes — for visual/layout QA without execute_js getComputedStyle snippets. |
| dispatch_pointer | Synthetic pointer/mouse gesture chains (click, dblclick, down, up, hover, drag) on the exact matched element, no retargeting. Element-relative offset for canvas hit-testing; drags interpolate moves on the element and doc |
| app_bridge | Calls helpers the host app registered via window.__MCP_BRIDGE__.register(name, fn, description) — list to discover, call to invoke (async supported, results serialized + truncated). The sanctioned path to app state inste |
| type_text | Types text into the page. Supports a fields array for bulk form fill, selector targeting, or typing into the focused element. Selects options in <select> dropdowns by value or label, and attaches files to <input type="fi |
| press_key | Presses keyboard keys the typing tools can't express: Escape, Enter, Tab, arrows, Backspace/Delete, F-keys, and modifier chords (cmd+a). Emulates default actions (Enter submits, Tab moves focus) unless the app prevents t |
| mouse_action | Non-click mouse actions: hover, scroll (by direction/amount/to element/to top/bottom), drag (start to end coordinates). |
| navigate | Webview navigation: goto (URL), back/forward (with optional delta), reload. |
| execute_js | Runs arbitrary JavaScript in the webview. Returns the last expression's value; promises are awaited. The universal escape hatch. |
| manage_storage | localStorage operations (get/set/remove/clear/keys) and cookie management (get/clear). |
| manage_window | Window control (list/focus/minimize/maximize/close/position/size/fullscreen), zoom, devtools, and webview state management. |
{
"mcpServers": {
"tauri-mcp": {
"command": "npx",
"args": ["tauri-plugin-mcp-server"]
}
}
}Configuration as documented by the project. Restart the client after saving.
| Variable | Description | Required |
|---|---|---|
| TAURI_MCP_IPC_PATH | Filesystem location the server is allowed to use. | Optional |
| TAURI_MCP_CONNECTION_TYPE | Configuration value read at startup. | Optional |
| TAURI_MCP_TCP_HOST | Endpoint or connection string the server talks to. | Optional |
| TAURI_MCP_TCP_PORT | Configuration value read at startup. | Optional |
| TAURI_MCP_AUTH_TOKEN | Credential the server authenticates with. | Yes |
Microsoft's official browser automation server — drive a real browser through the accessibility tree, no screenshots needed.
Industrial-strength web extraction — render, scrape, crawl and search entire sites into clean markdown.
The original Chromium automation reference server — simple, screenshot-driven browser control.
Give your coding agent the full DevTools toolbox: traces, network, console, heap snapshots and Lighthouse.
Puppeteer-powered browser control that drives pages from the accessibility tree instead of pixels.
Cloud browsers for AI agents — automation sessions that run in Browserbase's fleet, not on your machine.