Fake Windows 11 Blue Screen of Death_

A pixel-accurate recreation of the Windows crash screen that runs as a web page. It draws the sad face, the “Your PC ran into a problem” message, a climbing percentage, the QR code and a real stop code, and it fills the whole display in fullscreen. Nothing is installed, nothing is changed on the computer, and closing the tab ends it. It is a prank and a screensaver, not malware.

Click anywhere on the blue screen or press Enter Fullscreen. Once fullscreen, nothing is drawn on top of it — no hint, no exit button, no tooltip — so there is nothing on screen to give it away. Press Escape to leave.

Press Escape to exit fullscreen, then close or navigate away from the tab. Alt+Tab on Windows or Cmd+Tab on a Mac switches away from it too. Because it is only a web page there is nothing to undo and nothing to clean up.

Yes — fifteen real ones, including CRITICAL_PROCESS_DIED, IRQL_NOT_LESS_OR_EQUAL, MEMORY_MANAGEMENT, PAGE_FAULT_IN_NONPAGED_AREA and VIDEO_TDR_FAILURE. You can also let it cycle through them, which looks like a machine crashing repeatedly for different reasons. Pick one that suits the story you are telling: a graphics code on a gaming PC, a disk code on an old laptop.

Yes, with the speed slider. Slow is more convincing for a lingering prank; fast suits a quick reaction. With Loop forever on it restarts at 0% each time it reaches 100, so it never resolves — which is exactly how a machine stuck in a crash loop behaves, and why it holds up over a long lunch break.

Yes. Go fullscreen with looping on and it runs indefinitely, optionally cycling stop codes. People use it on office displays, conference stands, streams and Halloween setups. It never sleeps the machine on its own, so remember it will keep the display awake.

It renders anywhere with a browser, and fullscreen works on macOS, iPadOS and Android. It is far less convincing on a phone or a Mac for the obvious reason — those devices do not show Windows crash screens — so the prank really only lands on a Windows PC.

Yes. Use the black colour or the Windows 11 black preset. It is not just a recolour: it drops the sad face and the QR code and shows the plain sentence with the stop code and the faulting driver, which is how the crash screen has actually looked on Windows 11 since the mid-2025 redesign.

F toggles fullscreen, Space pauses the counter, R resets to defaults and Escape leaves fullscreen. Pausing is useful if you want the percentage frozen at a dramatic number for a screenshot.

Yes, freely — no watermark, no attribution required, no licence to read. It is a common prop in tutorials about troubleshooting, in security-awareness training, and in sketches. Recording it fullscreen gives you clean footage of a crash screen without having to crash anything.

A real BSOD ignores you completely. The mouse cursor disappears, the keyboard does nothing, Escape and Alt+Tab have no effect, and the machine reboots by itself once it has finished collecting the dump. A fake one is a web page: the cursor still moves, Escape exits, Alt+Tab switches away and the browser is still running underneath. If a blue screen reacts to your input, it is a prank.

Press Escape, then Alt+Tab. If a browser window appears behind it, you have been pranked and nothing is wrong with the machine. If the screen ignores every key and eventually restarts on its own, it was a real crash and worth investigating. Either way nothing has been installed by a page like this one.

Yes, and it is worth recognising. Tech-support scams show a fake crash or warning screen with a phone number to call, often with an alarming audio loop, and try to get you to hand over remote access or a payment. A genuine Windows BSOD never displays a phone number and Microsoft never asks you to call one. Close the browser, and never call a number that appears on a crash screen.

Effectively no. A web page runs in a sandbox and cannot write a kernel crash; the worst a malicious page normally manages is to hang or crash the browser itself. A crash screen that appears inside a browser window — with tabs and an address bar still visible — is a picture of a crash, not a crash.

Use it on someone who will find it funny, on a machine where nothing important is unsaved, and reveal it quickly. Do not use it on a colleague mid-deadline, on shared or production equipment, on someone who is anxious about technology, or anywhere it might trigger a real IT support call. A prank that costs someone an hour of work or a helpdesk ticket has stopped being a joke.

The screen Windows shows when the kernel hits an error it cannot safely recover from — a bug check, in Microsoft's terminology. Rather than continue in a state that might corrupt your data, the system halts deliberately, writes what it knows to a dump file and restarts. It is a controlled failure, not a random one, and the stop code is the reason it gave up.

The name of the bug check — the specific reason Windows halted. MEMORY_MANAGEMENT, IRQL_NOT_LESS_OR_EQUAL and CRITICAL_PROCESS_DIED are all stop codes. Each has a matching hexadecimal number, and it is the single most useful piece of information on the screen, because it narrows a vague “my PC crashed” into a specific class of fault.

A process Windows cannot run without has terminated unexpectedly — commonly a core system service killed by corruption, a bad update or failing storage. It is one of the most frequently seen codes and one of the least specific. Running sfc /scannow and then DISM /Online /Cleanup-Image /RestoreHealth to repair system files is the usual first step.

A driver tried to access memory it had no right to touch, at an interrupt level where that is not allowed. In practice it almost always means a faulty or mismatched driver, and often the one you installed most recently. Rolling back or updating the newest driver — network, graphics and storage drivers are the usual suspects — fixes the majority of cases.

Windows found its memory in a state that should be impossible. Genuinely faulty RAM is a real cause, so run the built-in Windows Memory Diagnostic or MemTest86 overnight. Unstable overclocks and XMP profiles cause it too, as do bad drivers corrupting memory they do not own — so a clean memory test does not fully clear the machine.

Something asked for data that should have been permanently in memory and was not there. The usual causes are failing RAM, a corrupt page file, or a driver referencing memory it already released. Test the memory first, and if the machine is otherwise healthy look at recently updated drivers and any anti-virus that hooks deeply into the system.

The graphics driver stopped responding and Windows could not reset it — TDR is Timeout Detection and Recovery. It points at the GPU driver, an unstable GPU overclock, overheating, or a power supply that cannot hold up under load. A clean driver reinstall using DDU, and checking temperatures under load, resolves most of them.

In rough order of how often they turn out to be the culprit: drivers (especially graphics, network, storage and anything from a device you just plugged in), faulty RAM, failing storage, overheating, unstable overclocking, and corrupt system files or a bad update. Malware and genuine Windows bugs happen but are far down the list.

Write down the stop code — everything else follows from it. Then ask what changed: a new driver, a new device, a Windows update, new RAM. Undo the most recent change first. If it crashes repeatedly, boot into Safe Mode, which loads a minimal driver set; if the crashes stop there, a third-party driver is almost certainly responsible.

In a memory dump, usually C:\Windows\MEMORY.DMP or a small one in C:\Windows\Minidump, and in the Event Viewer under Windows Logs → System, where a BugCheck event holds the stop code and parameters. Free tools such as NirSoft's BlueScreenView read minidumps and often name the offending driver file, which is far quicker than guessing.

Open the minidump with BlueScreenView or WinDbg and look at the module named in the fault — a .sys file usually identifies the vendor immediately (nvlddmkm.sys is NVIDIA, atikmdag.sys is AMD). For intermittent crashes, Driver Verifier stresses drivers until the guilty one fails, but enable it only when you can still get into Safe Mode, because it deliberately makes unstable systems crash more.

One isolated crash after a driver update usually means nothing. A pattern means something: the same stop code repeatedly points at one component; crashes that worsen with heat point at cooling or a failing part; crashes during disk activity point at the drive — check its SMART data. Random codes that change every time often mean failing RAM or an inadequate power supply.

Microsoft redesigned it in 2025 as part of the Windows Resiliency Initiative. The background became black, the sad face and QR code were removed, and the layout was cut back to one plain sentence with the stop code and the faulting driver — the goal being that the essential information can be read in about two seconds before the machine restarts. It rolled out on Windows 11 24H2 and 25H2 and later.

The colour is, on current Windows 11. The mechanism is not — it is the same bug-check screen doing the same job, just restyled, which is why people still call it a BSOD. Windows 10 and older Windows 11 builds still show blue, and the phrase is so embedded after thirty years that the new one gets called a blue screen anyway.

It depends on the target's machine. On Windows 10 or an older Windows 11 build, blue is what they would actually see. On a fully updated Windows 11 machine, black is now correct — and someone who knows their PC well may spot a blue screen as out of date. Blue is still more instantly recognisable to everyone else, which is why it remains the default here.

Largely inherited convention. Early Windows kernel error text was rendered in the default text-mode colours of the era, and blue with white text was both readable and already used elsewhere in Windows. It was never a decision anyone agonised over — which is part of why it could be changed thirty years later without much ceremony.

No single person designed the whole thing. Microsoft's Raymond Chen has written that he wrote the Windows 3.1 version, and Steve Ballmer wrote some of the text for an early message. The modern kernel bug-check screen dates from Windows NT, developed under Dave Cutler's team. The nickname “Blue Screen of Death” came from users, not Microsoft.

In Windows 8, in 2012, as part of a deliberate move away from the dense wall of hexadecimal that the NT-era screen showed. The reasoning was that the technical detail helped almost nobody in the moment, so the redesign led with a plain sentence and a face. It stayed through Windows 10 and early Windows 11, and was dropped in the 2025 black redesign.

Added in Windows 10, it encoded a link to a Microsoft support page. In practice it usually pointed at generic troubleshooting rather than anything specific to your stop code, which is a fair part of why it was removed. Its real value was that you could photograph the screen with a phone before the machine rebooted.

Yes — at COMDEX in April 1998, during a live demonstration of Windows 98, plugging in a scanner produced a blue screen in front of the audience. Gates quipped that it must be why they were not shipping Windows 98 yet. It remains the single most-replayed moment in the BSOD's history.

The name has meant three different things, which is why it causes confusion. Historically it described an OS/2 or early-Windows boot failure to a blank screen; more recently it described Windows booting to a black desktop with nothing on it, usually a graphics-driver or Explorer problem; and since 2025 it is the official colour of the Windows 11 crash screen.

Yes, with better manners. macOS shows a kernel panic — historically a multilingual “You need to restart your computer” panel, now usually a spontaneous restart with a report afterwards. Linux produces a kernel panic or an oops, dumping a stack trace to the console or the system log. Same underlying event: the kernel hit something it could not safely survive.

A deferred procedure call ran for far too long, or the system waited on a driver that never answered. On desktops the classic cause is an outdated storage controller driver — switching to the generic Microsoft AHCI/NVMe driver usually settles it — followed by old chipset and network drivers. A buggy SSD firmware produces it too, so check the drive vendor's updater before assuming Windows is at fault.

A call from a driver or system service raised an exception the kernel could not handle. The screen normally names the module, and the frequent offenders are all the same class of software: virtualisation, anti-cheat, VPN filters, screen recorders and third-party anti-virus. Remove the most recently installed low-level tool first, then update graphics and audio drivers.

A kernel data structure failed an integrity check — Windows found memory it depends on had been corrupted and stopped rather than trust it. The two common causes are a driver written for an older Windows version and failing RAM. Update or uninstall out-of-date drivers, then run a memory test overnight.

The Windows Hardware Error Architecture reported a fault the hardware could not correct itself, normally in the CPU, its memory controller or the power feeding them. Unlike most stop codes this one points at hardware rather than software: undo any overclock or undervolt, reset the BIOS to defaults, and check temperatures and the power supply. If it still happens at stock settings, a component is genuinely failing.

A device failed to complete a power transition — it did not wake, or did not sleep, inside the time Windows allows. It usually strikes on resume from sleep, and Wi-Fi adapters, GPUs and external drives are the usual culprits. Update those drivers, and in Device Manager untick Allow the computer to turn off this device to save power for the adapter concerned.

The memory-compression store hit an error it did not expect. In practice it is most often a storage problem — a failing drive, a loose cable or corrupt system files — though display drivers and aggressive anti-virus also trigger it. Check the drive's SMART health and run chkdsk /f /r before looking anywhere else.

Windows could not reach the disk it boots from. It appears after a failed update, a storage driver change, a BIOS switch between AHCI and RAID/Intel RST, or a dying drive. Putting the BIOS storage mode back the way it was explains a large share of them; if an update caused it, uninstall that update from the recovery environment.

One CPU core stopped responding to the others — an interrupt was never serviced. It points squarely at the processor: an unstable overclock, wrong voltages, a BIOS that needs updating, or a faulty chip. Reset the BIOS to defaults first; crashes that survive that on stock clocks are a hardware matter.

A kernel-mode program raised an exception nothing caught. The screen usually prints the responsible file in brackets — ntfs.sys, netio.sys or a vendor .sys — which is the fastest lead you will get from any blue screen. Update or remove whatever owns that file.

The filesystem driver found the volume in a state it could not make sense of, which normally means corruption on the disk rather than a bug in the driver. Run chkdsk /f /r on the drive and check its SMART data; a machine that produces this repeatedly is often running on a drive that is on its way out. Back up before troubleshooting further.

Every stop code has a number as well as a name, and older Windows versions showed only the number. 0x0000007B is INACCESSIBLE_BOOT_DEVICE, 0x0000000A is IRQL_NOT_LESS_OR_EQUAL and 0x0000001A is MEMORY_MANAGEMENT. The four parameters printed after it are debugging details meant for a minidump, not something you need to decode by eye.

Turn off automatic restart: System Properties → Advanced → Startup and Recovery → Settings, then untick Automatically restart. The crash screen will now stay up until you restart the machine yourself, which is long enough to photograph it. On the black Windows 11 screen the stop code and the faulting driver are both printed, so one phone photo captures everything you need.

Interrupt the boot three times — power on, then hold the power button while Windows is loading — and the Recovery Environment opens on the next attempt. From there: Troubleshoot → Advanced options → Startup Settings → Restart, then press 4 or 5. Safe Mode loads only Microsoft drivers, so if the crashes stop there, a third-party driver is almost certainly to blame.

A Windows 11 feature introduced alongside the black crash screen, as part of the Windows Resiliency Initiative. When a PC cannot boot, it enters the recovery environment, connects to the network and pulls a targeted fix from Microsoft without anyone touching the machine. It exists because of incidents where millions of PCs each needed a person standing in front of them.

Device Manager → the device → Properties → Driver tab → Roll Back Driver restores the previous version, if Windows still has it. If that button is greyed out, uninstall the device with delete the driver software ticked and reboot so Windows installs its own. For graphics drivers, DDU in Safe Mode gives the cleanest result.

Uninstall it: Settings → Windows Update → Update history → Uninstall updates, or from the recovery environment if the machine will not boot. Then pause updates for a week or two while the bad one is fixed. Crashes that start the day an update lands and stop when it is removed need no deeper diagnosis than that.

For memory, run Windows Memory Diagnostic (mdsched.exe) or, better, MemTest86 from a USB stick for several passes — faults often surface only after an hour. For storage, read the SMART data with a tool such as CrystalDiskInfo and run chkdsk /f /r. Reallocated or pending sectors are a reason to replace the drive, not to keep testing it.

Load and heat. That pattern points at the GPU driver, an unstable CPU or GPU overclock, memory running an XMP/EXPO profile the kit is not actually stable at, poor cooling, or a power supply that sags under a sudden draw. Run everything at stock with XMP off for a few days: if the crashes stop, it was never a Windows problem.

Get into the recovery environment by interrupting the boot three times, then work down the list: Startup Repair, uninstall the most recent update, System Restore to a point before it started, and Safe Mode to remove the suspect driver. Note the stop code from the loop first — INACCESSIBLE_BOOT_DEVICE points at storage or BIOS settings, while a named driver points at whatever was installed last.

No — the machine has already stopped, so nothing is running to take a screenshot. Photograph it with a phone, which is part of why the crash screen still fits the essentials on one page. If you need clean footage of a crash screen for a video or a slide, that is precisely what a fake one like this is for.

On 19 July 2024 a faulty CrowdStrike Falcon content update — Channel File 291 — caused an out-of-bounds memory read in a kernel driver and blue-screened roughly 8.5 million Windows machines worldwide, grounding flights and stopping hospitals and banks. CrowdStrike withdrew the update in about 80 minutes, but already-crashed PCs had to be repaired by hand, typically by booting into Safe Mode and deleting the offending file. It is the clearest demonstration there is that a blue screen is the symptom of one bad driver, not of a broken Windows.

It installs nothing and changes nothing. It is a single web page: HTML, CSS and JavaScript rendered in a sandboxed browser tab, with no downloads, no extensions, no permissions requested and no access to your files or system settings. Close the tab and every trace of it is gone.

It should not, and normally does not — there is no executable code, no download and nothing to quarantine. A small number of aggressive web-filtering products flag anything with “prank” or “fake error” in the page, which is a keyword match rather than a real detection.

No personal data. There is no account, no login, no file access and no upload. The site uses privacy-first analytics that count page views without cookies or personal profiles. Your stop-code choice and settings stay in your browser.

Once the page has loaded, yes — everything runs locally, so it keeps working with the connection off. That also means you can load it before going somewhere with no signal and still have it ready.

Completely free, with no account, no watermark, no time limit and no ads. It is one of the free browser tools at jasperbernaers.com, all of which run in your browser.