Observation

Screens, element trees and live views.

observe

GET /v1/devices/{deviceId}/observe

The agent’s look at the screen - screenshot reference plus a compact element list

The cheapest way for an agent to see a device. Elements come from the accessibility tree when the driver has one, or from on-device-free vision (OCR and control detection) when it does not. Secure fields and masked apps are redacted before anything leaves the engine.

Parameter	In	Required	Type	What it is
deviceId	path	yes	string	Stable device id such as ios:00008120-001A2B3C4D5E6F70 or android:R58M123ABC.
Silan-Lease	header	no	string	Lease token when the device is claimed.
screenshot	query	no	none, jpeg, png
maxLongEdge	query	no	integer
elements	query	no	none, interactive, all

Response	Means
200	Observation. Observation
4XX	Error (RFC 9457).
5XX	Error (RFC 9457).

MCP tools on this operation: observe.

screenshot

GET /v1/devices/{deviceId}/screenshot

Current screen as an image

Parameter	In	Required	Type	What it is
deviceId	path	yes	string	Stable device id such as ios:00008120-001A2B3C4D5E6F70 or android:R58M123ABC.
Silan-Lease	header	no	string	Lease token when the device is claimed.
format	query	no	png, jpeg, heic
maxLongEdge	query	no	integer
quality	query	no	number

Response	Means
200	Image. Coordinates in later calls may use space=image with the returned id.
4XX	Error (RFC 9457).
5XX	Error (RFC 9457).

MCP tools on this operation: screenshot.

getScreenshot

GET /v1/screenshots/{screenshotId}

Fetch a recent screenshot by id (kept in memory for 10 minutes unless part of a run)

Parameter	In	Required	Type	What it is
screenshotId	path	yes	string

Response	Means
200	Image.
4XX	Error (RFC 9457).
5XX	Error (RFC 9457).

getElements

GET /v1/devices/{deviceId}/elements

Full element tree

Parameter	In	Required	Type	What it is
deviceId	path	yes	string	Stable device id such as ios:00008120-001A2B3C4D5E6F70 or android:R58M123ABC.
Silan-Lease	header	no	string	Lease token when the device is claimed.
source	query	no	auto, accessibility, vision
visibleOnly	query	no	boolean

Response	Means
200	Element snapshot. ElementSnapshot
4XX	Error (RFC 9457).
5XX	Error (RFC 9457).

MCP tools on this operation: get_elements.

findElements

POST /v1/devices/{deviceId}/elements/find

Find elements matching a query

Parameter	In	Required	Type	What it is
deviceId	path	yes	string	Stable device id such as ios:00008120-001A2B3C4D5E6F70 or android:R58M123ABC.
Silan-Lease	header	no	string	Lease token when the device is claimed.

Request body: application/json (ElementQuery).

Field	Type	What it is
text	string
identifier	string
role	application, window, button, link, text, textField, secureTextField, searchField, image, icon, cell, list, switchControl, slider, picker, tab, tabBar, navigationBar, toolbar, alert, sheet, keyboard, keyboardKey, webView, notificationBanner, other
description	string	Free text like “the red Checkout button”; tree first, then vision.
match	exact, contains, regex, caseInsensitive
index	integer

Response	Means
200	Matches, best first.
4XX	Error (RFC 9457).
5XX	Error (RFC 9457).

MCP tools on this operation: find_elements.

streamMjpeg

GET /v1/devices/{deviceId}/stream.mjpeg

Simple live view for local tools (loopback only; the remote view uses WebRTC)

Parameter	In	Required	Type	What it is
deviceId	path	yes	string	Stable device id such as ios:00008120-001A2B3C4D5E6F70 or android:R58M123ABC.
fps	query	no	integer

Response	Means
200	multipart/x-mixed-replace stream of JPEG frames.
4XX	Error (RFC 9457).
5XX	Error (RFC 9457).