This guide focuses on how to spot anomalies in Animal Hospital: In this night-shift veterinary clinic, monsters disguised as animals are mixed in with the patients in line, and you must identify them before they enter. Mastering a consistent anomaly judgment process is the key to surviving shift after shift.
Below, we will break down the four-step inspection method for each patient, the specific tell-tale signs for each of the four inspection layers, the complete decision-making process, and which false red flags should not lead you to reject a patient. Whether you are a veteran of horror games or a newcomer to this genre, reading this will help you master how to spot anomalies immediately.
What is an anomaly? Why is identification the core of this night shift job?
Before we start breaking down the techniques, let's set the stage—understand the role of the "anomaly" in the game, and you will realize why how to spot anomalies is not just a bonus, but the bottom line for survival.
In Animal Hospital (a Roblox anomaly horror game), you play as a night-shift receptionist at a veterinary clinic. After dark, animal patients appear one by one at the reception window waiting to register; the problem is that monsters disguised as regular animals—known as anomalies—are mixed in. Your job is to identify them before they enter the clinic.
This is a co-op survival horror game for up to 30 players: you need to register normal patients and complete treatment mini-games in the medical room, while dealing with wandering enemies and unexpected events, all while keeping your Sanity above 0. The game has no end point, but continues infinitely in units of "Shifts" (nightly rotations)—the more shifts you can survive, the better. Therefore, mastering how to spot anomalies in Animal Hospital is the core survival skill for this job.
The gameplay of "anomaly horror" can be summarized in one sentence: FNAF meets "spot the difference," and you are the veterinary receptionist. It follows a slow-burn, oppressive psychological horror path, relying on atmosphere rather than frequent jumpscares. In other words, spotting anomalies relies on calm observation and point-by-point comparison, rather than speed or reflexes.
Core Principle: Better to reject wrongly than to accept wrongly
Before learning any specific techniques, engrave this mindset into your brain. It determines which way you should lean when you are unsure.
Remember this iron rule that applies to the entire game: The cost of incorrectly "rejecting" a normal patient is minimal—you only lose the rewards (Animal Coins and Cash) that this single patient would have brought; however, the cost of incorrectly "admitting" an anomaly is massive—it will enter the clinic and may turn into a Skinwalker (a lurking monster wearing a familiar skin), threatening the entire Animal Hospital and all your colleagues.
Therefore, when in doubt, always lean toward "rejecting".
This "asymmetric risk" dictates the correct mindset: you are not striving to "get every single one right," but rather to control the worst-case scenario. The most common mistake for beginners is to let their guard down because they fear losing rewards or because a patient "looks cute"; but once an anomaly is let in, the subsequent cleanup costs, Sanity loss, and team risks far outweigh those few Animal Coins.
Think of anomaly detection like airport security: it is better to stop one innocent traveler than to let one dangerous item through. Dropping that red shutter to keep a suspicious patient out is always the safer default option.
Standard inspection procedure for every patient (Four-Step Method)
For every patient that walks up to the window, use the same fixed procedure; do not skip steps based on your mood. Turning this into muscle memory is the foundation of how to identify anomalies at the Animal Hospital.
- ① Visual scan from head to tail
As soon as the patient arrives at the window, perform a complete head-to-tail visual inspection: head, eyes, limbs, fur patterns, and posture. Focus on whether there are extra or missing parts, misplaced facial features, or stiff/bizarre movements.
- ② Take a photo and compare it point-by-point with the subject
Press the button to take a photo and compare it point-by-point with the living patient at the window. Anomalies often reveal inconsistencies between the photo and the physical entity—colors, patterns, expressions, or even the number of parts may not match. Note: Taking a photo has a chance of triggering a cursed photo, which deducts 10 points of Sanity.
- ③ Check the paperwork
Review the registration paperwork: check if the name, species, and appointment information match the animal present. If the species is listed as a dog but something else arrives, or if the appointment doesn't match, these are clear red flags.
- ④ Check the CCTV if you are unsure
If you still have doubts, pull up the CCTV footage to take another look. Some flaws only reveal themselves on camera, while the patient may look perfectly normal at the window.
As long as even one confirmed red flag appears in any layer of the inspection, reject immediately—drop the shutter; you do not need to gather multiple pieces of evidence. The four inspection layers are an "OR" relationship, not an "AND" relationship. Conversely, only admit and register if all four layers are clean.
Detailed breakdown of the four inspection layers: Window, Photo, CCTV, Paperwork
The four-step method corresponds to four independent "inspection layers," each capable of catching different types of flaws. The table below provides a horizontal comparison, followed by a layer-by-layer breakdown. This layered approach is the most practical framework for how to distinguish anomalies.
| Inspection Layer | Where to check | Typical anomaly clues | Cost / Note |
|---|---|---|---|
| Layer 1: Window Visual | Observe the patient in real-time at the reception window | Extra/missing limbs, joint dislocation, misplaced eyes, hollow staring, inconsistent species traits, stiff posture | Free, fastest, always the first step |
| Layer 2: Photo Comparison | Compare point-by-point with the real-time subject at the window after taking a photo | Inconsistencies in color, pattern, expression, or number of body parts between the photo and the patient. | Risk of drawing a cursed photo, which deducts 10 Sanity per use. |
| Level 3: CCTV | Access camera footage to observe the patient. | Clues only visible on CCTV, while the patient appears normal at the window. | Requires leaving the window and takes time; used for verification when in doubt. |
| Level 4: Document Verification | Check name / species / appointment details. | Name, species, or appointment info does not match the animal present. | Free; mandatory for suspicious cases. |
Visual Window Inspection
- Too many/few limbs, or dislocated joints.
- Abnormal eye position or count, or an empty, hollow stare.
- Fur color or pattern does not match the species.
- Stiff movements, twitching, or strange posture.
Countermeasures：Free and fastest; perform this on every patient first. If a flaw is confirmed, drop the shutter immediately.
Photo Comparison Layer
- Color/pattern in the photo does not match the patient.
- Expression or number of body parts in the photo differs from the window.
- Photo contains details that shouldn't be there.
Countermeasures：Compare every detail against the patient; drawing a cursed photo deducts 10 Sanity, so use with caution when Sanity is low.
CCTV layer
- Flaws only visible on CCTV
- Normal at the window, but anomalous on CCTV
- Extra figures or distortions appearing on CCTV
Countermeasures：Check only when in doubt; requires leaving the window briefly, which slows down the pace, so no need to check every patient.
Document verification layer
- Name does not match registration
- Species does not match the animal present
- Appointment information does not match
Countermeasures：Free, must check if in doubt; mismatched information is a clear red flag, reject immediately.
There is a rumor circulating in the community about an 'Auditory / Sound' inspection layer—for example, an anomaly making sounds it shouldn't. However, this has not been officially confirmed and remains an unverified rumor, and community information is divided. Until it is proven, please rely on the four reliable layers: window, photo, document, and CCTV. Do not reject a patient based solely on strange sounds.
Decision-making flow: From visual inspection to admission or rejection
String the four-step method into a decision chain. Follow it for every patient so you won't miss a diagnosis or slow down the reception desk due to hesitation.
The standard decision chain is: Visual → Photo → Document → (if in doubt) CCTV → Decision. After each layer, ask yourself one question—did a confirmed red flag appear? If yes, stop further checks immediately, pull down the shutter, and reject. If no, proceed to the next layer.
Only after all four layers are cleared should you admit the patient at the reception desk and send them to the medical room for treatment via the mini-game. Note: CCTV does not need to be checked for every patient; it is a supplementary confirmation tool for when 'Visual / Photo / Document' leaves you unsure—it is time-consuming and requires leaving the window, so overusing it will slow down your pace.
If and only if: there are no visual anomalies from start to finish, the photo matches the patient perfectly, all document information aligns, and (if necessary) CCTV confirms everything is normal—only then is this a true patient. Register them with confidence. The standard for entry is zero red flags, not just a few.
Don't make mistakes: False red flags and decoys that are not anomalies
New players are often overly suspicious, misidentifying normal NPCs or environmental events as anomalies and rejecting them. The following look like red flags but actually aren't. Recognizing them will save you from many wrongful rejections and keep you from panicking.
Ratthew (friendly mouse NPC)
- Stares at the CCTV camera, which looks like a 'Staring' anomaly at first glance.
Countermeasures：Its head never turns beyond a natural angle—this is exactly how it differs from a 'Staring' anomaly. Don't panic or reject it just because it's looking at the camera.
Slime on the floor
- Suspicious slime appears on the clinic floor.
Countermeasures：This is an environmental trace, not a red flag on a patient, and cannot be used as a basis for rejecting a patient.
Fire in an occupied room
- Fire appears in a treatment room.
Countermeasures：Fire in a room occupied by a patient is an incident/event; treat it as such and don't misjudge the patient as an anomaly.
Ratthew's staring, slime on the floor, and fire in an occupied room—none of these three categories constitute evidence that a patient is an anomaly. They are either the normal behavior of friendly NPCs or environmental events. To determine the authenticity of a patient, always return to the four reliable sources of evidence: the window, photos, documents, and CCTV.
To distinguish between friendly characters and real threats, you can refer to our Character and Enemy Guide—it includes identification tips for NPCs like Ratthew, Barney, and Officer Duckman. For a systematic understanding of specific anomaly behaviors, see the Anomaly Guide.
The relationship between Sanity and identification efficiency
Identifying anomalies is not an isolated action; it is tightly bound to your Sanity resource. A good examiner is also a good accountant.
Sanity is a survival dashboard that must be kept above 0: if it hits zero, you won't survive the Shift. The problem is that the act of checking itself consumes Sanity—most notably in the photo comparison layer: when you take a photo, there is a chance of drawing a 'Cursed Photo,' which instantly deducts 10 Sanity.
This creates a practical trade-off: photos are a powerful tool for catching anomalies, but you can't spam them mindlessly. Use photos for precise comparisons when your Sanity is high; when Sanity is low, prioritize free visual checks and document verification, saving your precious Sanity for patients you are truly unsure about. Identification efficiency and Sanity management are two sides of the same coin.
Taking photos carries the risk of drawing a Cursed Photo, deducting 10 Sanity each time. This means that taking photos consecutively while at low Sanity could push you directly to zero. Treat photos as a key verification tool, not something to spam on every patient.
The cost of a wrong admission: How an anomaly becomes a Skinwalker
The reason we repeatedly emphasize that it's better to reject a real patient than to admit an anomaly is hidden in this section—the loss from a wrong admission is not linear, but systemic.
When you accidentally admit an anomaly, it won't stay at the registration desk. It will enter the clinic and may transform into a Skinwalker—a threat that wears a familiar face and wanders the hospital. It no longer just threatens the reward of a single patient, but the safety of the entire hospital and all your colleagues.
This is why you must identify anomalies before letting them in: once they are inside, the problem escalates from losing a few Animal Coins to the whole team having to spend Sanity and resources to deal with a wandering threat. Keeping it outside the shutter is always cheaper than dealing with it once it's inside.
The loss from a wrong admission is systemic: an anomaly inside can turn into a Skinwalker, threatening the entire team, far exceeding the Animal Coins / Cash reward of a single patient. This is the mechanical basis for why it is better to reject a patient than to accidentally admit an anomaly.
Common mistakes and easy-to-miss rejections for beginners
Learning from others' mistakes is easier than falling into them yourself. Below are the most common judgment errors made by new receptionists.
| Common mistakes | Why it's wrong | Correct approach |
|---|---|---|
| Rejecting Ratthew just for staring at the camera | That is normal behavior for a friendly NPC; the head does not rotate at a supernatural angle | Refer back to the four-layer evidence check; do not reject based on this |
| Admitting suspicious patients because they are cute or to avoid losing rewards | Admitting a patient might let in an anomaly; the cost is far greater than one reward | Always lean toward rejecting if in doubt |
| Taking photos of every patient | You might get a cursed photo, resulting in -10 Sanity | Prioritize visual inspection + documents when Sanity is low |
| Rejecting just because they sound weird | Auditory evidence is still unverified hearsay | Rely on the window / photos / CCTV / documents |
| Waiting for multiple red flags before rejecting | The four layers are an 'OR' relationship; one confirmed red flag is enough | Drop the shutter immediately upon confirming one red flag |
| Mistaking slime on the floor or room fires for patient anomalies | Those are environmental traces and random events | Handle them as events; they do not affect the validity of the patient |
Another hidden pitfall is over-reliance on CCTV: while it can catch flaws that only appear on screen, it requires you to leave the window and slows down your registration pace, which is not efficient during high-pressure shifts with many patients. Use CCTV as a final confirmation for doubts, not as a mandatory step for every patient.
Quick reference list: Know whether to admit at a glance
Condensing judgment into a table so you can decide at a glance when things get tense.
| Situation | Evidence strength | Decide |
|---|---|---|
| All four layers of inspection are clean | Zero red flags | Admit (register and send for treatment) |
| Any layer shows a confirmed red flag | ≥1 confirmed point | Reject (lower the red shutter) |
| A vague feeling that something is wrong | No confirmed evidence | Review photo + CCTV; if still doubtful, lean towards rejection |
| Ratthew staring at the camera / slime on the floor / room on fire | Evidence not related to the patient | Do not reject based on this; continue judging by the four layers |
| Photo is clearly inconsistent with the subject in some way | Photo layer confirmed | Reject |
| Document species / name / appointment do not match | Document layer confirmed | Reject |
Advanced: Turn the four-layer inspection into muscle memory
Only when the four-step method becomes instinct can you maintain your pace during a chaotic, high-intensity Shift with 30 patients.
The difference for advanced players isn't knowing what to check, but checking quickly and steadily. A fixed sequence (Visual → Photo → Document → CCTV if necessary) allows your eyes to develop a scanning rhythm, preventing missed judgments caused by looking around aimlessly; at the same time, it helps you manage your Sanity—save the high-cost photo and CCTV checks for truly doubtful cases, while using free visual and document checks to filter out the majority first.
When working in a team, you can divide tasks: someone focuses on window registration, while others rotate through CCTV and event handling, cross-checking patients when unsure. The game updates almost daily, and the specific manifestations of an anomaly may change with versions, so keeping the process framework stable and internalizing the methodology for Animal Hospital anomaly identification is more sustainable than rote memorizing the single traits of one specific anomaly.
Anomalies will increase or decrease with updates, but the framework of "four-layer inspection + reject on one red flag + better to reject wrongly" remains effective long-term. Master the method first, then go to the Anomaly Database to study the specific flaws of each anomaly for maximum efficiency.
FAQ
Click to collapse / expand
What exactly is an anomaly in Animal Hospital?
An anomaly is a monster disguised as a normal animal patient, waiting in the queue to enter the clinic. Your task is to identify them before registration and reject them using the red shutter.
How to quickly determine if a patient is an anomaly?
Follow a fixed four-step method for every patient: Visual scan → Photo comparison → Document verification → (if doubtful) Check CCTV. If even one confirmed red flag appears on any layer, reject immediately; only admit if all four layers are clean.
What happens if I wrongly reject a normal patient?
You only lose the reward for that patient (Animal Coins / Cash), which is a small price to pay. In contrast, wrongly admitting an anomaly could cause it to transform into a skinwalker and endanger the whole team, so always lean towards rejection when in doubt.
Does taking photos drain Sanity?
Yes. Taking photos has a chance to trigger a cursed photo, which drains 10 Sanity points. Use them sparingly when Sanity is low; prioritize free visual checks and document verification, saving photos for patients you are truly suspicious of.
Ratthew keeps staring at the CCTV camera, is it an anomaly?
No. Ratthew is a friendly NPC; its head will never rotate beyond a natural angle, which is the key difference between it and "staring"-type anomalies. Don't panic or shutter it just because it's looking at the camera.
Should I reject a patient because of slime on the floor or fire in the room?
No. Slime on the floor is an environmental trace, and fire in a room is a random event; neither constitutes evidence that a patient is an anomaly. Just handle them as events.
Do anomalies actually make sounds? Can I reject a patient immediately if I hear a strange noise?
The so-called "auditory layer" is currently still an unverified community rumor, and data is inconsistent. Do not reject based solely on sound; please rely on the four layers of reliable evidence: the window, photos, CCTV, and documents.
How serious is it to let in an anomaly?
Very serious. It may turn into a Skinwalker and roam the hospital, threatening the entire clinic and all your colleagues. The cost far outweighs the reward of a single patient—this is the basis for why it's better to reject a safe patient than to let one in.
What happens if Sanity hits zero?
Sanity must be kept above 0; if it hits zero, you won't survive the current Shift. Therefore, treat it as a core resource to manage, especially being mindful of the deductions from cursed photos.
Do I need to check CCTV for every patient?
Not necessarily. CCTV forces you to leave the window and slows down your registration pace. It should be used as a supplementary check when in doubt, rather than a mandatory step for every patient.
Which platforms does this game support?
PC, mobile, and tablet support are confirmed; console/VR are currently player rumors, with console expected to launch around June 26. Please refer to in-game announcements for specifics.