Random Animal Generator with Pictures, Facts and Red List Status

A random animal generator that shows real animals

Most animal pickers hand you a bare word. I wanted cards I could put in front of a class or a sketch group without double-checking every one, so this random animal generator draws from a hand-checked list of 310 animals in seven classes: mammals, birds, fish, reptiles, amphibians, insects and other invertebrates such as octopus, crab and starfish. Each card carries a picture, the scientific name, where the animal lives, its IUCN Red List status when there is a clear one, and a single fact that survives a look at the literature.

Filters for class, continent, habitat, first letter, difficulty and "threatened only" shrink the pool before the draw, and a live count and a small bar show how many animals are left in it. Three modes cover the jobs people actually bring here: a plain list, a drawing mode with a sketch timer, and a party mode for Pictionary-style drawing games and charades. Everything runs in your browser. Nothing is uploaded, there is no account, and the pictures are emoji drawn by your own device, not images fetched from somewhere else.

How to use it

1. Set the filters. Tick one or more classes, then narrow by continent, habitat, letter or difficulty if you need to. Watch the match count: if it drops to a handful, loosen something.
2. Choose how many. One to 24 cards. Repeats are off by default; if you ask for more than the filters allow, you get the whole pool and a note saying how many animals matched.
3. Generate. Press the button, or Ctrl+Enter (Cmd+Enter on a Mac), which also deals the next card in Draw and Party mode.
4. Flip a card. Tap or press Enter on a card to see its fact, habitat and a Red List scale, plus a "Learn more" link to the Wikipedia article, where you will find photos.
5. Keep or pass it on. Copy the names (with or without the scientific names), download CSV or TXT, print the cards eight to a page, or copy a share link that rebuilds the same list for someone else.

List mode

The everyday view: a grid of cards, or a compact list when you only need names. The last ten lists stay in a short history on this device, each with its seed and filters, so you can bring back yesterday's set exactly. Reset clears the filters and the results, and an Undo button stays for five seconds in case you pressed it by mistake.

Draw mode

One big card at a time, an optional twist such as "seen from behind" or "only basic shapes", and a sketch timer of 1, 2, 5 or 10 minutes with Start, Pause and Resume. The screen flashes when time is up, and a soft sound plays only if you switch it on. Next never repeats an animal until the whole pool has come up.

Party mode

The card stays hidden behind "Tap to reveal" so only the person drawing or acting sees it. You get a round timer of 30, 60 or 90 seconds, two to four teams with +1 and −1 buttons, and a deck that counts down what is left.

What is on each card

Tap a card to turn it over. Here is a typical card, front and back, field by field, so you know what you are reading:

Some cards deliberately have no badge. "Owl" or "Gecko" is a group of many species, and the IUCN only assesses species (sometimes subspecies or local populations), never a whole family or order. Several owls are doing fine and several are threatened, so a single "Owl · LC" badge would be wrong, and those cards say "group – many species" instead. Domestic animals such as the dromedary camel are labelled "domestic animal" for the same reason. Where I could not confirm a current status, I left the badge off rather than guess. Of the 310 animals, 45 are groups, 15 are domestic and 52 are species without a verified status, so 112 cards carry no badge.

Red List badges explained

The "threatened only" filter keeps CR, EN and VU, which is 74 of the 310 animals. The dodo is in the list too, with a grey EX badge. Statuses are not fixed, and three recent changes show how much they move:

• Green sea turtle: Endangered to Least Concern, after the global population grew by about 28% over several decades (IUCN update). The same update moved the hooded seal from VU to EN.
• European hedgehog: Least Concern to Near Threatened, with an estimated decline of about 30% over a decade from roads, development, intensive farming and pesticides (Cornwall Wildlife Trust).
• Migratory monarch butterfly: first listed Endangered, then changed to Vulnerable once the population models were re-examined (Monarch Joint Venture).

Random animals for drawing practice

I start every sketch session the same way: four animals at 2 minutes each, then one 10-minute study, 18 minutes in all. The short rounds are not about finished drawings. They force you to find the line of action and the big shapes before the clock runs out, which is exactly what you skip when you have all afternoon. A random pick matters here: left to choose, everybody draws a cat.

For the first rounds, set difficulty to easy. An elephant, a giraffe or a penguin has a silhouette you can read from across the room, so you can judge your proportions at a glance. Move to hard animals such as a tuatara or a walking leaf for the long study. Twists stop muscle-memory drawing: "mid-jump", "seen from behind", "one continuous line" or "in winter" make you think again about an animal you have drawn a hundred times. The full session plan, with a week of prompts by difficulty, is in random animal drawing prompts.

Animal Pictionary and charades without the arguments

Two things spoil animal drawing games: someone glimpses the word, and the same easy animals keep coming back. Party mode fixes both. The card stays covered until the drawer taps it, and the deck works like a real pile of cards: every animal in your filter comes up once before anything repeats, with a counter such as "7 of 40 left in the deck" so everyone can see it is fair. When the deck reshuffles, a notice tells you. Set difficulty to easy for young children, mix levels for a family table, and switch the label from Draw it to Act it out for charades. On a video call, the drawer can keep the generator open on a phone and draw on the shared screen.

Rules, timings, scoring and ready-made word lists by difficulty are in animal Pictionary and charades. Pictionary is mentioned here only as the name of the game; this site is not affiliated with its makers.

Animals A to Z, including the hard letters

The letter filter uses the first letter of the English common name, ignores a leading "The" and treats accented letters as plain ones, so vicuña counts as V. Every letter from A to Z has at least three animals, which matters most for the five letters where alphabet games usually stall:

A full alphabet table with three animals per letter, plus classroom activities, is in animals A to Z.

Is it really random? The dice behind every card

Each result starts with a fresh 32-bit seed from your browser's cryptographic random source, shown on screen in base 36 (seed 12345 shows as 9ix). The generator then takes your filtered pool in a fixed order (sorted by each animal's id), shuffles it with the Fisher–Yates method driven by a small, well-known generator called mulberry32, and deals the first cards. Because every step after the seed is deterministic, the random animal generator is reproducible. The same seed and the same filters always give the same animals, which is what makes share links work without any server: the link keeps the seed and filters after the # sign, a part of the address that browsers never send to a website.

With repeats off you cannot get the same animal twice in one result. With repeats on, every card is an independent draw, and repeats turn up sooner than most people expect:

For one particular animal the two settings barely differ: in 8 cards from the whole list of 310 it appears with a chance of 8 ÷ 310 = 2.6% without repeats and 1 − (309 ÷ 310)⁸ = 2.6% with them. The difference shows in games. Drawing one word at a time with replacement from 40 animals, you need on average 40 × (1 + 1/2 + … + 1/40) = 171.1 draws before you have seen every word. The party deck needs exactly 40.

Where the tool picks a number straight from the crypto source, it uses rejection sampling instead of a plain remainder. 2³² is not a multiple of 100 (it leaves 96 over), so a plain remainder would favour the smallest numbers very slightly. Values at or above 4,294,967,200 are thrown away and drawn again. These are the control cases the tool is tested against, each one computed separately in Python:

The randomness is the easy part. The real limit is the data: the list of 310 is curated by hand rather than pulled from a database, facts are kept short and conservative, and Red List statuses can change after any update. When it matters, follow the card's link and check.