Skip to main content
Glama

Robot Actions — Remote Device Control

webpage_click

Click an element in the device browser, by CSS selector or by a ref from a page-elements listing — works on BOTH iOS Safari and Android Chrome, auto-detected from the udid. Scrolls the element into view, waits for it to stop moving, and checks it is actually clickable — visible, non-zero-size, enabled, and not covered by anything — retrying until timeoutMs before it gives up. Then taps its center. On iOS the DEVICE taps its own screen at that position, so the page receives a real, fully trusted touch — pointerdown, touchstart, mousedown and click, exactly as from a finger — which also satisfies sites that gate on trusted input. On Android the tap is delivered as a TOUCH through the browser, so elements listening for pointer events respond as well as those listening for click. Falls back to mouse events, then a plain .click(), if a step is unsupported, and the response reports which was used via via. ERRORS rather than reporting success when the element never becomes clickable; pass force:true to skip the checks and dispatch at its coordinates anyway — which works for anything that still occupies a box, but NOT for a display:none or zero-size element, where there is no point to aim at and force fails too. When a selector matches several elements the first VISIBLE one wins, not the first in document order. This is COORDINATE-based. Android events are synthesized by the browser (isTrusted:false); iOS taps are real device input (isTrusted:true) whenever via comes back as "nativeTap".

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
refNoElement ref from a page-elements listing, e.g. "e12". Provide this or `selector`. A ref points straight at the element it was issued for, so it survives markup that shifted position, and it can address elements no CSS selector can reach from the top of the page.
udidYesDevice UDID / serial (iOS or Android)
forceNoSkip the clickability checks and dispatch at the element center regardless (default: false). Use only when the checks are wrong about a target you know is clickable — it restores the old behaviour, where a click on a hidden or covered element reports success without doing anything. Needs a center to aim at, so it still fails on a display:none or zero-size element; it does apply to visibility:hidden, opacity:0, pointer-events:none, disabled, off-viewport, and covered targets.
pageIdNoTarget page/tab id — auto-picked when omitted
socketNoAndroid only: abstract unix socket name (default: chrome_devtools_remote)
selectorNoCSS selector for the target element. Provide this or `ref`.
snapshotNoAppend a webpage_snapshot of the resulting page (actionable elements only) so you can see what the click changed without a second call (default: false)
timeoutMsNoHow long to keep retrying while the element is not yet clickable (default: 5000, max: 30000)
pointerTypeNoHow the tap is delivered. You rarely need this — the default is the best path each platform has (a real device tap on iOS, a browser touch on Android). "touch" is what a finger produces and drives elements that listen for pointer events, which many component libraries bind. On Android a touch also produces mousedown/mouseup unless a handler cancels pointerdown or touchstart — and when one does, click is suppressed too, which is when "mouse" is the way through. "mouse" sends mouse events only, and on iOS it opts OUT of the device tap in favour of the browser-level path, which delivers only mouseup+click and is not trusted. The response reports which was used via `via`.
platformVersionNoIgnored (kept for compatibility)

TDQS

A4.8/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries full burden and excels: it discloses scrolling, clickability checks, retry/timeout, fallback to mouse/.click(), error semantics, force behavior, first-visible-element selection, and platform-specific trust (iOS real tap vs Android synthesized). No contradictions.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The text is long but every sentence earns its place. It front-loads the basic purpose, then layers behavior, platform differences, fallback/error semantics, and parameter nuances without redundancy or filler.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given 10 parameters, platform-specific behavior, and no output schema, the description is remarkably complete. It covers event trust, fallback order, timeout behavior, failure cases, force limitations, and reports the `via` field, making it self-sufficient.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so baseline is 3. But the description adds substantial meaning beyond the schema: it explains how ref survives layout shifts, how force behaves with hidden/covered elements, how pointerType affects event delivery and trust, and how selector multiple matches resolve.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description opens with 'Click an element in the device browser, by CSS selector or by a `ref`' — a specific verb and resource. It distinguishes this from sibling tools like web_click or device_tap by explicitly stating it works on both iOS Safari and Android Chrome and is coordinate-based.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description gives strong context: when to use (device browser on iOS/Android), when to pass force, and which pointer type to choose. It clearly indicates this is for device browser vs native app taps, but it does not explicitly name alternative tools or give a 'use X instead' exclusion.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.

TDQS

B3.1/5.0
Disambiguation2/5

The set contains near-identical duplicate families: web_* and playwright_* expose ~15 pairs of the same desktop-grid-browser operations (web_get_text/playwright_get_text, web_reload/playwright_reload), and screenshot/log/network/mock capabilities each have 5-8 entry points (device_screenshot vs android_mjpeg_screenshot vs ios_screenshot vs ios_fast_screenshot vs web_screenshot vs webpage_screenshot vs session_screenshot). Many individual descriptions carefully draw boundaries (devtools vs traffic, HID vs session), but an agent cannot reliably distinguish web_* from playwright_*, and ios_screenshot/ios_fast_screenshot/ios_mjpeg_screenshot blur together.

Naming Consistency2/5

The prefix scheme is broken: Android functionality is split arbitrarily between android_* and device_* (device_screenshot vs android_mjpeg_screenshot), the desktop browser gets two parallel prefixes (web_* and playwright_*), and verbs vary across equivalents (device_navigate_url vs web_navigate vs ios_safari_navigate). session_* uses bare verbs (session_url, session_back), and the same concept gets different names (ios_clipboard_get_hid vs ios_get_pasteboard; device_screen vs ios_orientation).

Tool Count1/5

333 tools is an extreme count by any measure — far beyond the 50+ threshold — and much of the bulk is duplicative (the web_*/playwright_* pairs alone double ~15 slots) or out-of-scope for a device-control server (TestRail, Jira, AzDO, agent memory, secret variables, feedback). Even granting that remote device control + test automation is a broad domain, this surface will devastate agent context budgets and is impossible to navigate coherently.

Completeness4/5

The core device-control and test-automation domain is remarkably thorough: Android and iOS each have full interaction, app-lifecycle, file, network/proxy, performance, crash, accessibility, recording, and replay coverage, with CRUD lifecycles for flows, suites, app uploads, TestRail cases, and visual-review baselines. Minor gaps exist at the margins — Jira/AzDO lack update/transition/comment operations, and iOS cannot open/close tabs — but the central workflows have no dead ends.

Resources