A face memory game in your browser can mean two very different things, and the puzzle-style version built around code-drawn illustrations asks you to memorize and identify five fictional profiles for exactly 1,000 points. Unlike camera-driven identification apps that scan a real face and return a name, this category of game gives you one illustrated profile to study for a moment, then asks which of three alternatives matched it. The Face Memory Test is a clean example of that style: five rounds, code-drawn SVG profiles, a deterministic scoring path, and a fully local interface that never opens your camera or uploads anything. That distinction matters because the phrase face recognition game online often pulls up biometric software, GitHub experiments that hook into a webcam, or research demos that print a match probability. Those tools are interesting, but they require a working camera, often a model download, and they treat the result as an identity claim. A visual memory face game flips the question: instead of asking whose face is this, it asks which drawing did you just see. That keeps the activity short, deterministic, and accessible on any device with a browser, and it runs in a single HTML page with no setup, no install, and no permission prompts.
What a Face Memory Game Means in a Browser
The honest answer to what a face recognition game is depends on which kind of recognition you want. Biometric recognition games take a photograph or a webcam frame and return a guess about who the person is. Memory-style recognition games take a brief visual stimulus and ask whether you can pick it back out from a small set of alternatives. Both are valid, but they solve different problems and need different tools.
The Face Memory Test sits firmly in the second camp. The fifteen profile illustrations are original geometric SVG drawings assembled from a small authored vocabulary. There is no photo, no stock image, no AI-generated person, and no identity record behind any of them. Each one is a fixture, named Alder through Opal, that exists only inside the game. The recognition task is purely visual recall: you saw drawing X two seconds ago, and you now need to choose drawing X again from among three options.
That setup keeps the result auditable. Every target and every alternative in every round is a literal fixture in the game's code, so the same round always presents the same three options in the same positions. There is no shuffle, no random portrait, no adaptive difficulty, and no hidden score adjustment. The run is fully reproducible from the moment you press start.
Play the Face Memory Test in Three Steps
- Study the original illustrated target during the two-tick Reveal, or choose Recall now when ready.
- After the target unmounts, select the matching fictional profile with a large button or keys 1 through 3.
- Match all five profiles before a second distinct miss to finish with exactly 1,000 points.
Each round is short and explicit. The Reveal phase mounts one profile illustration and no answer buttons, then counts down through two visible one-second ticks. You can wait for both ticks to pass on their own, or you can press Enter, Space, or the Recall now control to hide the profile earlier if you have already memorized it. When the target unmounts, three option buttons appear in a single horizontal row. Pick the one that matches the illustration you just saw, either by clicking the large button or by pressing 1, 2, or 3 on the keyboard.
A clean match adds 200 points and immediately opens the next round's Reveal. After five clean matches, the run ends at exactly 1,000 points. If you miss once, you can keep playing and recover, but the rules for what counts as a second miss are strict, which is why a quick walk-through of the controls and the run state helps before you start.
The Fifteen Profiles Are Code-Drawn, Not Photographed
The fifteen faces in the game are not pulled from any dataset. They are built from a small, authored SVG vocabulary that the game composes in code at load time. Each profile is a combination of one head outline, one hair shape, one brow shape, one eye shape, one nose shape, one mouth shape, and optional cheek marks or freckles. The vocabulary is fixed, so the same fifteen faces appear in every session in the same order.
| Profile feature | Authored options |
|---|---|
| Head outline | Round, oval, square |
| Hair shape | Five original silhouettes |
| Brow shape | Three variants |
| Eye shape | Three variants |
| Nose shape | Three variants |
| Mouth shape | Four variants |
| Optional marks | Cheek marks or freckles |
These are puzzle features, not measurements of real people. The named profiles Alder through Opal are fictional labels used to reference fixtures inside the code. They do not represent a person, demographic group, diagnosis, emotion label, or biometric template. That is why the game can claim it never recognizes a face, never compares an upload, never estimates identity, and never infers a sensitive attribute: the underlying illustrations have nothing to recognize.
Controls, Keyboard Shortcuts, and Round Flow
The pointer path uses three large illustrated buttons in a single row. Each button carries a visible textual label, so you always know which option you are about to pick. The buttons have a visible focus state, and the status text announces the current round, your running match count, and your error count without relying on color alone.
The keyboard path covers the same flow without a pointer. During the Reveal phase, press Enter or Space to hide the target early. After the three options mount, press 1, 2, or 3 to choose directly. You can also press Left or Right to move the visible selection from button to button, then press Enter or Space to confirm it. Up and Down deliberately do nothing because the option tray is one row, so vertical keys would only confuse the focus order.
| Phase | Mouse or touch | Keyboard |
|---|---|---|
| Reveal | Wait for two ticks, or tap Recall now | Enter or Space to hide the target |
| Recall | Tap one of three large option buttons | Press 1, 2, or 3 directly |
| Selected-button route | Tap the focused button | Left/Right moves, Enter/Space confirms |
| Up / Down arrows | Not used | No-op (single-row tray) |
| Restart | Tap the shared Restart control | Activate the Restart button |
A few actions are deliberately atomic no-ops, which means they do nothing and do not change the run state. Pressing an unknown identifier, choosing an option that belongs to another round, moving the cursor outside the three-button tray, or trying to act during the wrong phase all leave the round exactly as it was. That includes duplicate wrong picks on the same profile: if you choose the same wrong option twice in a round, the error count stays at one. Only a second distinct wrong profile triggers the run-ending deadlock.
Scoring, Miss Recovery, and When the Run Ends
The score is fully deterministic. A clean match is worth 200 points, and there is no time bonus, no streak multiplier, no random portrait, and no hidden adjustment. The arithmetic is direct: 200 points per clean match × 5 rounds = 1,000 total points. Five clean matches therefore end the run at exactly 1,000 points. That simple calculation is the entire scoring formula, which makes the final number easy to verify by hand against any completed run.
The recovery rule is one of the easiest things to get wrong on a first read. The game allows one distinct wrong profile across the entire run. If you choose the same wrong profile again in the same round, that pick is deduplicated and the error count does not increase. Only when you pick a second different wrong profile, in any round, does the run close and disable further answers.
| Event | Effect on score | Effect on run |
|---|---|---|
| Correct match | +200 | Advance to next target Reveal |
| First distinct miss | No change | Continue the round, attempt recovery |
| Same wrong profile again | No change | Deduplicated, no-op |
| Second distinct miss | Run closes | Answer buttons disabled, run frozen |
| Five clean matches | Exactly 1,000 | Run completes |
The fixed option positions across the five rounds are 2, 1, 3, 2, and 1 when read as the labeled buttons the player sees. Knowing that the correct answer sits in position 2 in round one, position 1 in round two, and so on is useful only as a sanity check, because the game already enforces the correct path. The position list is included here so you can audit any run against the published fixture.
Why Nothing Leaves Your Browser
The whole experience is local. There is no camera request, no file upload, no account, no API call, no machine-learning model, and no cookie-based identity. The illustrations are SVG with semantic foreground, surface, accent, and border variables, so they render correctly against the site's themes without downloading a single image asset. Each illustration has a stable view box that reserves its dimensions, so the layout never shifts while you study the profile.
The only state that persists in your browser is your shared best completed score. You can clear it by clearing site data in your browser. Everything else, including which profile you saw last and which option you just picked, lives only inside the active run.
Two extra interface features round out the controls. A shared Restart control returns to the first target and restores the two-tick countdown for a fresh attempt. A double-press of the Escape key opens a simulated spreadsheet boss screen for a quick cover, and pressing Escape twice again resumes the run in place.
Other Visual Memory Challenges to Try
If you enjoy this style of recognition game, several sibling tools share the same deterministic, code-drawn approach. The Visual Memory Test works with five fixed grid patterns instead of faces. The Pattern Memory Game shows a 3×3 pattern, then asks you to pick the exact hidden record from four options. The Position Memory Game stores three shapes on a 3×3 board and asks you to restore their hidden positions. The Shadow Matching Game compares drawn object specimens against shuffled silhouettes. Each one finishes on an exact 1,000-point route, none requires a camera, and none makes claims about your memory ability.
The shared appeal is the same as in the Face Memory Test: a short, reproducible challenge that you can audit down to the position of every button and the exact value of every point. That combination is what turns a casual face recognition game into a tool you can verify rather than a result you have to trust.