Use Playwright MCP’s browser_take_screenshot tool to create the image. To find it reliably, pass a filename, such as artifacts/login.png, then open that path under your workspace root. If you leave out filename, look in the MCP server’s configured output directory for a timestamped file named like page-{timestamp}.png, page-{timestamp}.jpeg or page-{timestamp}.webp.
Find the file: explicit path or automatic filename
The location depends on whether you supplied filename to browser_take_screenshot:
- You supplied a filename: open that path relative to the workspace root. For example, if you used
artifacts/login.png, look forartifacts/login.pngin the workspace. - You omitted the filename: check the directory configured for Playwright MCP output. The generated name follows the
page-{timestamp}.{png|jpeg|webp}pattern. The output directory is configurable, so there is no single automatic-file location to assume for every setup.
These are two different location rules: a relative, explicitly supplied filename is resolved from the workspace root; an automatically named screenshot goes under the configured output directory. If you cannot find a file, first check which case applies, then verify the path or output-directory setting for the MCP server you started. The screenshot tool’s options are documented in the Playwright MCP Screenshots reference.
Capture and locate a screenshot step by step
- Start or configure the Playwright MCP server. The official setup uses the
@playwright/mcppackage; see the Playwright MCP getting-started guide for setup. The exact way you start and configure an MCP server depends on your client and configuration, so use the appropriate client setup rather than assuming one universal command. - Navigate to the page you want to capture. Use your MCP client’s browser-navigation workflow to load the target page and wait until it is in the state you want to document. For a login screen, for example, decide whether the screenshot should show the initial form, an error state, or the page after a successful login.
- Choose the output location before capture. For a predictable path, provide an explicit filename such as
artifacts/login.png. If you do not provide one, note the configured output directory so you know where to look for the timestamped file afterward. - Call
browser_take_screenshot. Choose whether you need the visible viewport, an element, or the full scrollable page; then set format and scale if the defaults are not appropriate. - Open the saved image. With an explicit filename, locate that workspace-relative path. Without one, inspect the configured output directory and match the generated
page-{timestamp}name and extension.
For repeatable scripts and review workflows, an explicit filename is usually easier to manage: you know what path to open, and you do not need to identify the newest timestamped file. If you want each capture to be unique, use distinct filenames for distinct states rather than assuming an automatic name will be written in the same place as a relative filename.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstall#1 Best Overall
Choose viewport, element, or full-page capture
The screenshot tool can capture a page viewport, a chosen element, or the complete scrollable page. Pick the scope based on what someone needs to inspect, not just on how much of the page will fit in an image.
| Capture scope | How to request it | Useful for | Constraint |
|---|---|---|---|
| Viewport | Call browser_take_screenshot without a target or fullPage. |
A specific visible state, such as a form, menu, or modal that is already on screen. | It represents the viewport, not the entire scrollable document. |
| Element | Pass an element reference or selector using target. |
A particular component, chart, dialog, or other region when the rest of the page is irrelevant. | Do not combine target with fullPage. |
| Full page | Set fullPage to true. |
A long page that needs to be reviewed from top to bottom in one capture. | It captures the full scrollable page and cannot be combined with target. |
An element capture and a full-page capture answer different questions. Use target when the goal is a close view of one region; use fullPage when the entire document matters. If you need both a page overview and a focused component image, take two captures rather than trying to combine mutually exclusive options.
Set format, filename, and resolution
Choose an image format
Playwright MCP supports png, jpeg, and webp. You can set type directly. If you leave type unset, the format is inferred from the filename extension; if there is no extension to infer from, the default is PNG. For example, use a filename ending in .webp when you want the extension to make the format clear, or set type explicitly if you need to control it independently of the filename.
Keep the filename extension consistent with the chosen format so the saved file is easy to identify later. Do not assume that an extensionless automatic name will be JPEG or WebP: the documented default when no type can be inferred is PNG.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Choose CSS or device scale
The scale option controls screenshot sizing:
cssis the default and produces CSS-pixel sizing.deviceuses device-pixel sizing for a higher-resolution capture.
Use the default when a CSS-pixel-sized image suits your inspection or documentation need. Use device scale when you specifically want a higher-resolution image. The scale choice is independent of capture scope: you can select the scale while choosing a viewport, element, or full-page screenshot, subject to the rule that target and fullPage cannot be used together.
Make the filename useful
A filename is part of the workflow, not just a cosmetic detail. A name such as artifacts/checkout-error.png tells a teammate what state the image represents and where to retrieve it. If a path contains directories, confirm those directories are available in the workspace you are using; if the expected file is not there, check the exact path supplied and whether your MCP client is showing the same workspace root as the server.
Example capture choices
The exact way your MCP client displays tool inputs varies. The values below show the options to pass to browser_take_screenshot; they are not a shell command.
Save a predictable viewport screenshot
filename: "artifacts/login.png"
type: "png"
scale: "css"This requests a PNG at CSS-pixel sizing and gives you a known workspace-relative path. Open artifacts/login.png under the workspace root after the call.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errorsSave the complete scrollable page
filename: "artifacts/pricing-page.webp"
type: "webp"
fullPage: true
scale: "device"This requests a full-page WebP at device-pixel sizing. Since no target is present, the full-page option is not in conflict with an element capture.
Capture one element
target: "main .checkout-summary"
filename: "artifacts/checkout-summary.png"
type: "png"The target may be an element reference or a selector. Use the selector that identifies the element you want in the current page. Do not set fullPage for this call.
Use snapshots to interact; use screenshots to inspect visually
A screenshot is an image of what the page looks like. It is useful for visual review, saving evidence, and checking layout, but it is not the way Playwright MCP identifies controls for its interaction tools. Playwright’s documentation puts the distinction plainly: “Screenshots are for looking at, not for acting on — use browser_snapshot for refs to interact with.” See the Screenshots reference.
For interaction, use browser_snapshot to inspect the accessibility tree and obtain refs, then pass those refs to action tools. If a large snapshot makes a control difficult to locate, browser_find can help find text. After navigation, take a fresh snapshot before reusing refs: refs belong to the current page state and can become stale when the page changes. The snapshot workflow is described in the Playwright MCP snapshots documentation.
A practical sequence is to use the snapshot to identify and operate a control, then take a screenshot to record the visual result. If the page has navigated or changed state, refresh the snapshot before selecting another ref. Do not use a screenshot image as a substitute for the accessibility information needed to target actions.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Troubleshoot a missing or unexpected screenshot
The screenshot file is missing
- Cause: You are checking the wrong directory. An explicit relative filename is rooted at the workspace; an automatically generated file is written to the configured output directory.
- Fix: Check whether the call included
filename. If it did, look for that exact path under the workspace root. If it did not, inspect the MCP server’s output-directory configuration and search for a file namedpage-{timestamp}with a supported extension.
The file has a different extension or format than expected
- Cause: The format may have been inferred from the filename, or may have defaulted to PNG if neither the filename nor
typespecified one. - Fix: Set
typetopng,jpeg, orwebp, or use a filename with the matching extension. Check both the filename and the option passed to the tool when diagnosing a mismatch.
A full-page and element capture cannot be requested together
- Cause:
fullPagecannot be combined withtarget. - Fix: Make separate calls: one with a
targetfor the element and another withfullPage: truefor the whole document.
An interaction ref no longer works
- Cause: The page changed after the snapshot, so its refs may no longer describe the current state.
- Fix: Call
browser_snapshotagain after navigation or other page changes, then use the fresh ref for the action. A screenshot can help you see the current appearance, but it does not refresh interaction refs.
The screenshot shows the wrong part of the page
- Cause: The call captured the viewport when you needed the whole page, or targeted an element when you intended to capture a different scope.
- Fix: For the full scrollable document, set
fullPagetotrueand omittarget. For a specific component, pass that element’s selector or current ref astarget. For the visible view, omit both.
Or skip the browser setup
If your goal is a website screenshot from an application or script rather than a local Playwright MCP capture, ScreenshotNeo can return an image or PDF from one GET request. For example, this cURL command saves a WebP capture of Stripe. Replace the example URL and use your API key:
ScreenshotNeo API documentation
curl -G "https://api.screenshotneo.com/v1/shot" -d access_key=YOUR_API_KEY --data-urlencode url=https://stripe.com -o shot.webpIt is a different workflow from Playwright MCP: send a request to the screenshot API instead of starting a browser tool session. ScreenshotNeo removes cookie and consent banners, newsletter popups, and chat widgets before capture; each of those steps can be turned off. Bot checks or CAPTCHAs, blank pages, timeouts, failed loads, and cache hits are not billed, and responses include X-Page-Verdict and X-Billed headers to identify the result and billing status. Its MCP server offers take_screenshot, get_page_info, and capture_pdf for Claude, Cursor, and other MCP clients.
The free plan includes 1,000 shots per month with no card. Paid plans start at $5 for 3,000 shots; the other listed plans are Growth at $15 for 15,000, Pro at $39 for 60,000, Scale at $99 for 250,000, and Business at $249 for 1,000,000. Yearly billing gives two months free, and every feature is available on every plan. The API also supports PNG and JPEG, among its other capture options.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Sign up for ScreenshotNeo’s free plan to get 1,000 screenshots a month with no card.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
