OpenWand MCP Server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| get_selected_textA | Read the text the user currently has highlighted on their desktop. Works best when the selection is in the app the user last used. If this returns no selection, ask the user to copy the text and call get_clipboard instead. |
| get_clipboardA | Read the user's current clipboard text. Reliable on every platform and needs no window focus — the fallback when get_selected_text returns nothing. |
| get_active_windowA | Report the window the user is working in (title, app, URL when it is a browser), skipping the assistant's own window. |
| read_browser_pageA | Read the text of the page open in the user's visible browser window (Chrome, Edge, Firefox, Safari...), even when the browser is not focused. Returns the URL plus the page text. |
| take_screen_snipA | Take a screenshot of the user's primary monitor and return it as an image. Use when the user asks about something visible on their screen. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 5 tools
Each tool captures a distinct source of user context: selected text, clipboard, active window, browser page, and screenshot. There is no overlap in purpose, and the descriptions clearly differentiate when to use each (e.g., selected text vs clipboard fallback).
All five tools follow the verb_noun pattern with snake_case: get_selected_text, get_clipboard, get_active_window, read_browser_page, take_screen_snip. The verbs (get, read, take) are appropriately descriptive and consistent in style.
Five tools is well within the ideal 3-15 range and perfectly scoped to the server's purpose of reading user desktop state. Each tool earns its place with no redundancy or bloat.
The set covers all primary ways an agent can acquire user context from the desktop: selection, clipboard, active window info, browser content, and visual screenshot. No obvious gaps exist for the stated purpose.