Skip to main content
Glama

Screenshots of a website

take_screenshots
Destructive

Capture screenshots of a website's key pages in desktop, mobile and full-page sizes by opening it in a headless browser, hiding cookie banners and floating boxes, and saving organized files.

Instructions

Use it when the user wants screenshots of a site, App Store or Google Play screenshots, Product Hunt gallery images, or a web app manifest's screenshots. Opens a website in a headless browser, picks the pages worth showing (home, pricing, features and the like), hides cookie banners, chat bubbles and other floating boxes, and saves desktop, phone and whole-page screenshots in one folder. sheet.png, every page on one picture, is the picture to show first; SCREENSHOTS.md lists each page with why it was picked, its files and its headings. url and page are pages of one public site: a #route, like /app/#/plants, opens that view of a web app, nothing behind a login is reached, and above one page it starts at the site's root. clean hides everything fixed on the screen but a top bar and a side bar, so an app's own tab bar too. iphone is 1320x2868, the App Store's 6.9 inch size; mobile is 1080x1920, for Google Play phones. For the Google Play feature graphic and Chrome Web Store tiles, use make_launch_stills instead. For a view behind a click, use record_demo_video with steps. The screenshots playbook (makersmcp_playbook) has the whole job. (easyscreenshotkit 0.1.3, on this machine.)

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "url": "<url>"
    },
    "name": "take_screenshots"
  }
}

Input Schema

NameRequiredDescriptionDefault
pwaNoAlso save screenshots for a web app manifest in pwa/: each page's first screen at 1920x1080 and on a phone at 1080x1920, at most 8 in all, checked against Chrome's rules for its richer install dialog, with the manifest entries in pwa/manifest-screenshots.json.
urlYesThe website, like https://linear.app, or a page of it. A #route in it, like https://acme.app/app/#/plants, opens that view of the app; another fragment, like #pricing, is left out.
hideNoCSS selectors of more elements to hide, like .promo-bar, when something still covers a page.
langNoThe site's version in this language, when it has one: a language code like de, tr or pt-BR.
pageNoPages of the same site to take too, like /pricing, or views of a web app by their #route, like /app/#/plants; #/plants alone is that view of url. Each counts toward pages and is taken even past it, and with pages 1 only url and these are taken. Another fragment, like #pricing, is left out.
cleanNoSet to false to take the pages as they come; hide still hides what it names. Otherwise cookie banners, chat bubbles and everything else fixed on the screen are hidden, except a bar along its top and a tall bar down a side, so an app's own bottom tab bar is hidden too.
pagesNoHow many pages, the home page included (default 5). Above 1 the site's root comes first, then url when it is a deeper page, then each page, then pages from the site's menu. With 1, only url is taken, though a page that is not of the site still stops the run.
sizesNoAny of desktop (3200x1800), mobile (1080x1920, a Google Play phone screenshot), iphone (1320x2868, the App Store's 6.9 inch size), full (the whole page at 1x), mobile-full (the whole phone page at 2x), or WxH from 240 to 3840 a side, like 1270x760 for a Product Hunt gallery image. Each takes an optional @scale from 0.5 to 4, and each size goes in once (default desktop, mobile and full).
themeNoThe site's light or dark look, or both, which adds a -dark twin of every picture (default light). A site without a dark theme looks the same in both.
out_dirNoFolder to save into, absolute or relative to /app. Default: easyscreenshotkit-<host>, where <host> is the site's host without www or the port. A run replaces what an earlier run of this tool left there once it has ended well, and one that fails or is canceled leaves it as it was: give each version its own out_dir. The folder must be empty or hold an earlier run of this tool: any other folder stops it.
wait_secondsNoHow long to wait for the result, counted from the start of the call, before returning a job_id to wait on. Some clients end a call after 60 or 240 seconds; 45 is safe in all of them.
reduced_motionNoAsk the site for its reduced-motion version.
timeout_secondsNoGive up after this many seconds (default 300). A run takes two or three minutes, up to five on very long pages, and a slow site can take 600.

Output Schema

NameRequiredDescriptionDefault
kitYes
docsNoThe notes the text shows, and where each one is.
nextNoWhat to do next, as the Next line says.
toolYes
filesNo
folderNoWhere the kit saved its files.
job_idYes
passedNoFor checks: whether the check passed.
statusYes
continuedNoTrue when this call continued the job of an identical call whose answer did not reach the client, instead of starting another.
exit_codeNo
last_lineNoWhat the kit was doing last.
expires_atNoWhen the download links stop working.
next_actionNo
elapsed_secondsYes

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed19 schema fields changedv0.2.0
    • addedInput schema / additionalProperties
      Added value: +false
    • changedInput schema / properties / clean / description
      Previous value: -"Set to false to take the pages as they come, banners and bubbles included."New value: +"Set to false to take the pages as they come; hide still hides what it names. Otherwise cookie banners, chat bubbles and everything else fixed on the screen are hidden, except a bar along its top and a tall bar down a side, so an app's own bottom tab bar is hidden too."
    • changedInput schema / properties / hide / description
      Previous value: -"CSS selectors of more elements to hide."New value: +"CSS selectors of more elements to hide, like .promo-bar, when something still covers a page."
    • addedInput schema / properties / lang / anyOf
      Added value: +[
      +  {
      +    "minLength": 1,
      +    "type": "string"
      +  },
      +  {
      +    "type": "number"
      +  }
      +]
    • changedInput schema / properties / lang / description
      Previous value: -"The site's version in this language, when it has one, like de."New value: +"The site's version in this language, when it has one: a language code like de, tr or pt-BR."
    • removedInput schema / properties / lang / minLength
      Removed value: -1
    • removedInput schema / properties / lang / type
      Removed value: -"string"
    • changedInput schema / properties / out_dir / description
      Previous value: -"Folder to save into, absolute or relative to /app. Default: easyscreenshotkit-{host}."New value: +"Folder to save into, absolute or relative to /app. Default: easyscreenshotkit-<host>, where <host> is the site's host without www or the port. A run replaces what an earlier run of this tool left there once it has ended well, and one that fails or is canceled leaves it as it was: give each version its own out_dir. The folder must be empty or hold an earlier run of this tool: any other folder stops it."
    • changedInput schema / properties / page / description
      Previous value: -"Pages to take too, like /pricing."New value: +"Pages of the same site to take too, like /pricing, or views of a web app by their #route, like /app/#/plants; #/plants alone is that view of url. Each counts toward pages and is taken even past it, and with pages 1 only url and these are taken. Another fragment, like #pricing, is left out."
    • changedInput schema / properties / pages / description
      Previous value: -"How many pages, the home page included (default 5)."New value: +"How many pages, the home page included (default 5). Above 1 the site's root comes first, then url when it is a deeper page, then each page, then pages from the site's menu. With 1, only url is taken, though a page that is not of the site still stops the run."
    • changedInput schema / properties / pwa / description
      Previous value: -"Also save screenshots for a web app manifest in pwa/: each page's first screen at 1920x1080 and on a phone at 1080x1920, checked against Chrome's rules for its richer install dialog, with the manifest entries in pwa/manifest-screenshots.json."New value: +"Also save screenshots for a web app manifest in pwa/: each page's first screen at 1920x1080 and on a phone at 1080x1920, at most 8 in all, checked against Chrome's rules for its richer install dialog, with the manifest entries in pwa/manifest-screenshots.json."
    • changedInput schema / properties / sizes / description
      Previous value: -"Any of desktop, mobile, iphone, full, mobile-full or WxH, each with an optional @scale (default desktop, mobile and full)."New value: +"Any of desktop (3200x1800), mobile (1080x1920, a Google Play phone screenshot), iphone (1320x2868, the App Store's 6.9 inch size), full (the whole page at 1x), mobile-full (the whole phone page at 2x), or WxH from 240 to 3840 a side, like 1270x760 for a Product Hunt gallery image. Each takes an optional @scale from 0.5 to 4, and each size goes in once (default desktop, mobile and full)."
    • changedInput schema / properties / theme / description
      Previous value: -"The site's light or dark look, or both (default light)."New value: +"The site's light or dark look, or both, which adds a -dark twin of every picture (default light). A site without a dark theme looks the same in both."
    • changedInput schema / properties / timeout_seconds / description
      Previous value: -"Give up after this many seconds (default 300)."New value: +"Give up after this many seconds (default 300). A run takes two or three minutes, up to five on very long pages, and a slow site can take 600."
    • changedInput schema / properties / url / description
      Previous value: -"The website, like https://linear.app"New value: +"The website, like https://linear.app, or a page of it. A #route in it, like https://acme.app/app/#/plants, opens that view of the app; another fragment, like #pricing, is left out."
    • changedInput schema / properties / wait_seconds / description
      Previous value: -"How long to wait for the result before returning a job_id to wait on. 45 is safe for clients that stop a tool call after 60 seconds."New value: +"How long to wait for the result, counted from the start of the call, before returning a job_id to wait on. Some clients end a call after 60 or 240 seconds; 45 is safe in all of them."
    • addedOutput schema / properties / continued
      Added value: +{
      +  "description": "True when this call continued the job of an identical call whose answer did not reach the client, instead of starting another.",
      +  "type": "boolean"
      +}
    • addedOutput schema / properties / docs
      Added value: +{
      +  "description": "The notes the text shows, and where each one is.",
      +  "items": {
      +    "additionalProperties": false,
      +    "properties": {
      +      "name": {
      +        "type": "string"
      +      },
      +      "path": {
      +        "type": "string"
      +      },
      +      "url": {
      +        "type": "string"
      +      }
      +    },
      +    "required": [
      +      "name"
      +    ],
      +    "type": "object"
      +  },
      +  "type": "array"
      +}
    • addedOutput schema / properties / next
      Added value: +{
      +  "description": "What to do next, as the Next line says.",
      +  "type": "string"
      +}
  2. Added

TDQS

A4.7/5.0
Behavior5/5

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

Annotations already declare destructiveHint=true and openWorldHint=true, but the description adds substantial context beyond them: cookie banners and floating boxes are hidden, nothing behind a login is reachable, a run takes 2-3 minutes, and out_dir is replaced on success but preserved on failure or cancellation ('give each version its own out_dir'). It also discloses the async contract — a job_id is returned after wait_seconds — which the annotations do not cover.

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

Conciseness3/5

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

It is front-loaded well — the use case and primary output come first — and most sentences carry operational detail. However it is one dense block of run-on clauses, and several phrases duplicate the schema verbatim (size dimensions, defaults, theme behavior), which is wasted space in a 13-parameter tool.

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?

For a 13-parameter, open-world, destructive, long-running capture tool, the description covers what an agent needs: entry points, what gets hidden, what lands in the folder (sheet.png first, SCREENSHOTS.md as index), the async job_id/timeout path, and the out_dir collision rule. An output schema exists, so it correctly does not attempt to enumerate return fields.

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

Parameters4/5

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

Schema coverage is 100%, so the baseline is 3, and the description does add real meaning: url/page are pages of one public site, #routes open web-app views, non-route fragments are dropped, pages>1 starts at the site root, and clean keeps only top and side bars. It loses a point for restating schema text nearly verbatim (iphone 1320x2868, mobile 1080x1920, default sizes) rather than adding new information.

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?

States a concrete verb and resource ('opens a website in a headless browser ... saves desktop, phone and whole-page screenshots') and scopes it to specific intents (site screenshots, App Store/Google Play, Product Hunt gallery, PWA manifest). It also explicitly distinguishes itself from make_launch_stills and record_demo_video, so an agent can route without opening the schema.

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

Usage Guidelines5/5

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

Gives explicit trigger conditions ('when the user wants screenshots of a site, App Store or Google Play screenshots, Product Hunt gallery images') plus two named alternatives with the conditions that select them: make_launch_stills for feature graphics/store tiles and record_demo_video with steps for content behind a click. It also points to makersmcp_playbook for the full job.

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