capture

Capture and recordingFull

Capture into the library: a window, a rect, a display or every display, SuperCapture itself, or through the browser extension a tab's page, a presentation or a bookmark. A new item lands in folder (a rel path from library_find folders, '' the save root), else where the user is looking (state's capture_destination), named by the naming template; name fills its name slot (else AI rename names it, when the user has it on). open shows it in its editor, waiting for the load. The result names it (id, path). To look without saving, screen look; a page by its address, capture_url.

Needs the Full level, which is the default for any client holding the bearer token.

Parameters

Name	Type	Required	Description
action	string one of "window", "region", "display", "screen", "self", "page_visible", "page_full", "presentation", "bookmark"	yes	What to do. window: A window: window_id from screen list, or title (exact, else a substring) with owner to tell same-titled windows apart. (window_id, title, owner, folder, name, open, delay) region: A rect of the screen: [x, y, width, height] in global screen points, as screen list and screen locate report them. (rect, folder, name, open, delay) display: One display: an id from screen list (default the main display). (display, folder, name, open, delay) screen: Every display as one picture, each where it sits. (folder, name, open, delay) self: SuperCapture's main window. (folder, name, open, delay) page_visible: A browser tab's visible page, through the SuperCapture extension: the active tab, or the one tab_title / tab_url names. (tab_title, tab_url, folder, open) page_full: A browser tab's whole page, scrolled and stitched by the extension (fixed headers and sticky elements handled). (tab_title, tab_url, folder, open) presentation: Every slide of a presentation open in a tab, as one slide deck: Google Slides, DocSend, Canva, Gamma, Speaker Deck and Issuu are recognised; use_generic advances any other with the arrow keys until two slides match. (tab_title, tab_url, use_generic, name, folder, open) bookmark: A page open in a tab as a bookmark, as the extension's Bookmark button makes one: its first screen, title and article text, filed in Reader mode's library (folder, else the reading folder open there, else Reader). A feed page with no post of its own is refused. (tab_title, tab_url, folder, open)
window_id	integer	no	window: the window's id from screen list.
title	string	no	window: or its title, case-insensitive (an exact match wins over a substring).
owner	string	no	window: the app that owns it, as screen list reports owner.
rect	integer[]	no	region: [x, y, width, height] in global screen points.
display	integer	no	display: its id from screen list (default the main display).
tab_title	string	no	page_visible / page_full / presentation / bookmark: a tab title substring (default the active tab).
tab_url	string	no	Or a tab URL substring: a tab whose own host or path it names is preferred; include a scheme (https://...) to require an exact origin.
use_generic	boolean	no	presentation: advance with the arrow keys, for a site not recognised (reveal.js, Prezi, PowerPoint Online...).
folder	string	no	Where it goes: a rel path, '' the save root (default: the folder open in the gallery; a bookmark: the reading folder open in Reader mode, else Reader).
name	string	no	The name slot of its file name.
open	boolean	no	Open it in its editor and wait until it has loaded (default false).
delay	integer	no	window, region, display, screen, self: seconds to wait before taking it, as the capture timer does (default 0): time for a menu or hover state to be on screen.

Example

A JSON-RPC call to POST http://localhost:21516/mcp with the bearer token, filling in the required parameters:

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "capture",
    "arguments": {
      "action": "window"
    }
  }
}

See the MCP server overview for how to connect a client and where the token lives.