File LR-LEARNSubject pump.fun coinsCost free

Rug school.

From zero to reading a coin like a pro in 7 short levels: real pictures from our lab, one idea at a time, and a quick check after each level. Basic first, advanced after.

§01Rug school7 levels · about 2 minutes each

Your path

7 short levels, one picture at a time. Tap a level to start. Tap any green word to see what it means.

0 of 7 levels done
Advanced from here
★The finish lineRug school graduateFinish all 7 levels to get the badgeYou made it. Badge unlocked.

Level 1: The basics

A new coin every 2 seconds 46,678new coins, 21 Oct ≈32a minuteour recorder

Thousands of coins a day

Anyone can launch a coin on pump.fun in seconds. Most of them are gone within hours.

Names are free to copy QBTCQBTC contract addresscontract address FJFT…zBpump3bkm…qkDhc coin Acoin B 2 real coins, launched 2 s apart.Only the CA tells them apart.

Copycat coins

Two real coins called QBTC launched two seconds apart. The contract address (CA) is the only real ID, so always check it.

Every buy pushes the price up launch ≈ $3k full at ≈ 85 SOLgraduates ≈ $43k SOL put into the curve →price

The bonding curve

Every coin starts on a bonding curve: each buy pushes the price up, each sell pulls it down. Market cap is price × supply.

Fewer than 3 in 100 graduate 425 graduatedof 15,149 coins with 10+ trades (moved to PumpSwap)our data, 22 Oct

Graduation

When about 85 SOL is in the curve, the coin graduates to PumpSwap. Very few get there.

Quick check

Two coins are both called QBTC. How do you find the right one?

Quick check

What happens when a coin's bonding curve fills up?

Level 2: Read a coin

Little liquidity, big price drop small poolbig drop big poolsmall drop same sell size on both sides

Liquidity

Liquidity is the money on the other side of your trade. The less there is, the more your own sell pushes the price down.

A few wallets held most of it top 10 walletseveryone else 62% 35% caution line Over the line, a few wallets can swingthe price hard, both ways. our data: Kimichi, 22 Oct

Who holds it

If the top 10 wallets hold over 35%, a few people can move the price hard, both ways. LabRat shows it as a caution.

Your slippage is a gap bots can use price shown you pay your limit Low limit: the trade can fail.High limit: bots can make you pay it all.

Slippage

Slippage is how much worse than the shown price you accept. Too low and your trade fails; too high and bots can make you pay more.

Fees eat small trades0.01 SOL22.5%0.05 SOL6.5%0.1 SOL4.5%0.5 SOL2.9%1 SOL2.7%5 SOL2.5%fees on one buy + one sell, by trade sizeour cost sheet

Fees eat small trades

On a 0.01 SOL buy and sell, fees take about 22%. From 0.5 SOL up it is under 3%. A priority fee is part of it.

Quick check

You buy 0.01 SOL of a coin and sell at the same price. Roughly what do fees take?

Quick check

The top 10 wallets hold 60% of a coin. What does that mean?

Level 3: Spot the rug

17 wallets in the first 2 seconds first 2 s 17 wallets took79%of the supply normal buyers, laterour data: CC, 22 Oct

Bundled launch

Wallets working together buy a big part of the supply in the first seconds, then sell into the people who buy after them.

QBTC: insiders sold in seconds 0s1m2m38m 1234 1 dev sells at 4s2 bundle sells at 23s 3 top ≈ $19k at 75s4 back at launch, 38m our data: real trades from our recorder, 23 Oct

A real one: QBTC

QBTC was a bundled launch. The dev sold 6.1 SOL four seconds in, the bundle sold at 23 seconds, and the price was back at launch 38 minutes later.

One dev, 347 coins in a day 1 DEVwallet 347coins launchedin 24 hours 339dumped Nearly all named "eCash".our data, 23 Oct · each square ≈ 3 coins

Serial launcher

One dev wallet launched 347 coins in 24 hours and dumped 339 of them. Check the dev before you buy.

Back at launch price = dead launch price top $5.9k now $2.9k Everyone who bought after launch is down.our data: a Rug or not coin, 22 Oct

Dead at launch price

A coin doesn't need to crash 90% to be dead. Back at its launch price with no buyers means everyone after launch lost.

93 in 100 die within 3 hours 93 dead7 not dead Red flags raise the odds.No flags does not mean safe.our data: 8,180 coins, 21 to 23 Oct

The honest odds

93 in 100 normal coins were dead within 3 hours. Red flags make it likelier. No flags never means safe.

Quick check

8 wallets bought 20% of the supply in the first 2 seconds. What is it?

Quick check

A scan shows no red flags. Is the coin safe?

Level 4: Tools of the trade

Read the trades, not just the line PAID 312 1 price · 2 trades, long red = big sells3 a paid badge is an ad, not a check

Chart pages

DexScreener and similar sites show the chart and every trade. Look at who is selling, not just the line. A paid profile or boost only proves someone paid.

Joined bubbles = one group red, joined: one teamgrey, loose: other holders

Bubblemaps

Every holder is a bubble. Bubbles joined by lines moved tokens between each other, so one group shows up as a cluster.

Each tool sees different things RugCheck"Good"no risks found checks tokenrights, liquidity LabRat scan3 flagsbundle, dev dumped checks who boughtfirst and the dev Same coin, QBTC. Use more than one tool.our check, 23 Oct

Each tool sees different things

RugCheck rated QBTC Good with no risks found. Our scan found a bundle and a dev who dumped. Check more than one tool.

One tap from a scan LabRatscan BubblemapsRugCheckDexScreeneryour app open in… on every scan, plain links

One tap from a scan

Every LabRat scan has Open in… with that exact coin on Bubblemaps, RugCheck, DexScreener and your trading app.

Quick check

On Bubblemaps, 5 of the top holders are joined by lines. What does it suggest?

Quick check

A coin has a paid DexScreener profile and a boost. What does that prove?

Level 5: The trader playbook

3 seconds late is too late 0s1s2s3s botsyou All 150 exit rules lost money at 3 s late.our test

Snipers

Snipers are bots that buy in the launch second and pay for priority. Buying 3 seconds later lost money under all 150 exit rules we tested.

Copying 20 seconds late 0s10s20s top wallet buys then sells you copy The top 1% sold 5 to 15 s after buying.our data, Oct 2026

The copy-trading trap

The top 1% bought in the launch second and sold 5 to 15 seconds later. If you copy them 20 seconds late, you become their exit liquidity.

A trend spikes, then fades fast 323 in one hour new coins named "cyber", hour by hourour data: real hourly counts

Narrative waves

A narrative spikes and fades within hours. When a word is trending everywhere, the early buyers are leaving.

Kimchi wave: 1,518 coins in 36 h 237 in one hour21 Oct 20:0023 Oct 07:00 72% of the busy ones: the dev already soldour data, times in UTC

A wave you can watch: kimchi

When a trader called Kimchi was in the news, 1,518 new coins with kimchi in the name launched in 36 hours, up to 237 in one hour. Of the 538 that got real trading, the dev had already sold in 72%.

The call came after the pump callers buy here call posted Check the post time against the chart.

Calls

Many callers bought before they posted. Compare the time of the post with the chart.

Claimed vs traced on-chain claimed$40Mnever verified traced on-chain≈$6Mlinked wallets source: Bubblemaps on X, 22 Oct 2026

Check the claim, not the screenshot

A famous trader's $40M profit on $TRUMP was repeated for months, but no one ever showed the wallet. Bubblemaps traced the wallets it could link to him and found about $6M (there may be others). A wallet is proof; a screenshot is a story.

Cut losers fast 28%of normal traderswere in profit the bestsold losers fast the restheld and hoped They didn't win more often. They lost less.our data, Oct 2026

What actually helps

Only 28% of normal traders were in profit. The best didn't win more often: they cut losing trades faster. Set a stop loss before you buy.

Quick check

You copy a top wallet's buy 20 seconds later. What usually happens?

Quick check

What did the most profitable normal traders do differently?

Quick check

Someone posts a screenshot of a huge profit. What actually proves it?

Level 6: Dev dark arts

How a wallet gets traced 1start with a wallet you know2follow where SOL and tokens moved3linked wallets = one cluster4add up what they bought and sold Anyone can do it: the chain is public.

Tracing a wallet

Start from a wallet you know, follow where SOL and tokens moved, and linked wallets show up as one group. Bubblemaps and Solscan let anyone do it for free.

Insiders buy in the launch block one block, in this order CREATE INSIDERBUY INSIDERBUY … No one can get in between.You start after them.

Jito bundles

A Jito bundle runs the coin's creation and insider buys in order in one block. Nobody can get in between.

Many small holders, one owner SOLone source Each holds a little (1.5 to 5%).Together they are one big holder.

Split wallets

The supply is spread over many fresh wallets with a small share each, so the top 10 looks normal. Bubblemaps shows them as one group.

Busy chart, no real people buysell same size, same 2 wallets Fake activity to climb trending lists.

Volume bots

Volume bots buy and sell the same small amounts over and over: fake activity to climb trending lists.

A bot on both sides of your trade BOT BUYSfirst YOU BUYpay more BOT SELLSto you one blockHigh slippage is the bot's profit.

Sandwich (MEV)

A bot buys just before you and sells right after. You pay more; it keeps the difference. High slippage invites it.

Mayhem coins: AI trades, not people normal coin1B supplypeople tradeprice ≥ launch Mayhem coinsupply not 1BAI agent tradescan drop below 21%of busy new coins were Mayhemour data, 21 to 23 Oct

Mayhem mode

In Mayhem mode an AI agent trades the coin. Lots of trades, few real people. LabRat marks these coins with a caution.

Quick check

Many wallets each hold 2 to 4% and Bubblemaps shows them joined. What is it?

Quick check

Why does a bot buy right before you and sell right after?

Level 7: Stay safe

Never type your seed phrase Verify your walletenter your 12-word phrase CONTINUE SCAM No real site, bot or support needs it.

Your seed phrase

No real site, bot or support person ever needs your seed phrase. If anything asks for it, close it.

Verify bots drain wallets "Safeguard" verify bot paste this command toprove you are human VERIFY SCAM Real groups never ask you to run anything.

Fake verify bots

Telegram bots that ask you to paste a command or connect your wallet to verify are how wallets get drained.

Free tokens are bait your wallet SOL USDC FREE token: "claim" You never bought it. Hide it, never click.

Free tokens

Tokens you never bought, with a claim link, are bait. Hide them and never click.

Same start, same end, wrong middle your address7xKp…mR4tQ…Lz8V scammer's look-alike7xKp…b7Hd2…Lz8V They send you a tiny amount so it showsin your history. Copy from Receive instead.

Look-alike addresses

Scammers send tiny amounts from look-alike addresses, hoping you copy the wrong one from your history.

Quick check

A Telegram bot asks you to paste a command into your computer to verify. What do you do?

Quick check

Where should you copy your own wallet address from?

§02Before you buytwo minutes, every time

The 7-step checklist

Even with every box ticked, most coins still go to zero. This list keeps you out of the worst traps. Tap a green word to see what it means.

  1. Check the CA. Copy it from pump.fun or the real chart, never from a reply or a DM.
  2. Scan it. Bundled launch? Dev bought 5% or more? Read "what usually happens". Coin scan →
  3. Check the dev. Is this coin number 5 today? Dev check →
  4. Look at the holders. Linked bubbles mean one group. Bubblemaps ↗
  5. Read the chart, not the screenshots. Who bought before the hype? DexScreener ↗
  6. Size and exit first. Only money you can lose, and decide where you sell before you buy.
  7. Protect the wallet. Trade from a separate wallet, never share the seed phrase, never sign on a site you did not type yourself.
§03Dictionary38 words · like a wiki

Look up a word

Every word from the levels, in plain language. Search, or tap a green word to jump between them.

Start here

CA (contract address)

The coin's address on Solana, its one true ID. Anyone can copy a name or a picture, nobody can copy the CA. Always check the CA, not the name.

Market cap (mcap)

Price times supply. A pump.fun coin has 1 billion tokens, so a price of $0.00001 is a $10k market cap. It says how big a coin is, not how much money you could actually sell for.

Liquidity

The money on the other side of your trade. With little liquidity, your sell pushes the price down a lot. On pump.fun the bonding curve is the liquidity; after graduation it sits in a PumpSwap pool.

Bonding curve

pump.fun's built-in market. No order book: a formula sets the price. Every buy raises it, every sell lowers it. A coin starts around a $3k market cap.

Graduation (migration)

When about 85 SOL has gone into the curve, the coin moves to PumpSwap and trades in a normal pool. Under 3% of active coins get there.

DEX, AMM, PumpSwap

A DEX is an exchange run by code, with no company holding your coins. An AMM is a pool of two tokens where the price comes from the ratio between them. PumpSwap is pump.fun's own AMM for graduated coins.

Slippage

How much worse than the shown price you accept. Set it too low and your trade fails; set it too high and bots can make you pay much more (see sandwich).

Priority fee

An extra fee so your transaction gets in faster. Fast apps add it for you. On small trades, fees eat a big share: see the cost sheet.

Holders and top 10

The wallets that hold the coin, and how much the 10 biggest hold together. The bonding curve and pools are not counted. A big top 10 means a few wallets can move the price hard, both ways.

Dev (creator)

The wallet that launched the coin. Many devs buy their own coin at launch and sell it into the first buyers. Check what the dev did before you buy: Dev check.

Rug (rug pull)

Insiders sell everything and the price collapses. On pump.fun nobody can pull the curve's money, so a rug means the dev and friends dump their bags into buyers.

Bundle

Several buys packed into the launch by wallets working together, often in the same block as the creation. They own a big part of the supply before anyone else can buy.

Sniper

A bot that buys in the first second, paying high fees to be first. You can't win that race by hand.

Whale

A wallet making big trades. On our live tape, any trade of 1 SOL or more is marked as a whale.

Volume

How much was traded. High volume looks exciting, but bots can fake it (see wash trading).

Ape, bag, bagholder

To ape is to buy fast without checking. Your bag is what you hold. A bagholder holds a coin that already fell.

Exit liquidity

The late buyers that insiders sell to. If you buy after a big pump on a bundled coin, you are probably their exit.

Narrative (meta)

The theme everyone is copying right now, like a word, an animal or a news story. Waves rise and die within hours: Narrative radar.

KOL and calls

A KOL is an influencer. A call is a post telling followers to buy. Many callers are paid, or bought before posting.

Paid profile and boost

On DexScreener anyone can pay to add a coin's info or to boost it in the trending list. It only means someone paid. See them on our launch log.

CTO (community takeover)

The dev left, and a group took over the coin's socials and promotion. Sometimes real, sometimes the same insiders with a new story.

Take profit, stop loss

Selling part when you are up, and selling when it falls to a level you chose before buying. Plans made before the trade beat feelings during it.

DYOR, NFA

Do your own research, not financial advice. Everyone says it; few do the first part.

Seed phrase

The 12 or 24 words that control your wallet. Whoever has them has your money. No real site, bot or support person ever needs them.

LARP

Pretending: a made-up story or an unproven claim, like a celebrity link or a giant profit, used to pump a coin or a reputation. Ask for the wallet.

Go deeper

Jito bundle

A group of transactions sent together through Jito, a Solana block service, and run in order in the same block. Launch bundles use it so the insider buys land right after the creation.

MEV and sandwich

A bot sees your trade coming, buys just before you and sells right after. You pay a worse price; they keep the difference. High slippage makes it easy for them.

Wash trading (volume bots)

Buying and selling to yourself to fake activity and climb trending lists. Lots of trades, few real people.

Insider cluster (split wallets)

One group spreads its supply over many fresh wallets so no single holder looks big. Bubblemaps shows them as linked bubbles.

FDV

Fully diluted value: price times the maximum supply. On pump.fun all tokens exist from the start, so FDV equals market cap. On other tokens it can be much bigger.

Mint and freeze authority

The right to create more tokens, or to freeze wallets. pump.fun's own program handles its coins, so this matters most for tokens made elsewhere: check them on RugCheck.

LP tokens

A receipt for money in a pool. Whoever holds the LP tokens can take that liquidity out; if they are burned, nobody can. Check this on tokens outside pump.fun.

Creator fees

pump.fun pays the coin's creator a small share of every trade (0.30% on the curve when we checked). Some devs earn from volume, not from the coin going up.

Mayhem mode

A pump.fun launch mode where an AI agent trades the coin and the supply is not the usual 1 billion. Trade counts and prices behave differently, so read them with care.

Honeypot

A token you can buy but not sell. Rare on pump.fun because pump.fun's own program runs the trades, but common with tokens made elsewhere.

Drainer

A site that empties your wallet once you sign its transaction. It often hides behind "claim", "verify" or "mint" buttons.

Address poisoning

Scammers send you a tiny amount from an address that looks almost like yours, hoping you copy it from your history later.

Wallet tracing

Following where SOL and tokens moved between wallets to find who is really behind a coin or a claim. Bubblemaps and Solscan do it for free.

§04Toolsnot affiliated · no referral links

Tools traders use

What each site is for. Bookmark the real addresses: scammers buy ads for look-alike sites.

§05Techniquesgood and bad

What traders and devs do

Marked Fair play Grey area Manipulation, with our own numbers where we have them.

What traders do

Fair play

Check before buying

Scan the coin, look at the dev's other coins, open the holders on Bubblemaps, then look at the chart. Two minutes that skip most traps.

The checklist ↓
Fair play

Small size, exit planned

Decide how much you can lose and where you sell before you buy. Very small trades lose to fees: on a 0.01 SOL buy and sell, fees take about 22%; on 0.1 SOL about 4.5%.

Cost sheet →
Fair play

Cut losers fast

In our data, only 28% of normal-looking traders were in profit, and the best ones did not win more often: they cut losing trades faster and lost less on them.

Findings →
Grey area

Sniping

Bots buy in the launch second, paying high fees and Jito tips to be first. Legal, but a race of machines. Our test: buying every new coin 3 seconds after launch lost money under all 150 exit rules we tried, once real delays and failed transactions were counted.

Findings →
Grey area

Copy trading

Following the wallets that win most. The catch: the top 1% bought in the launch second with 3 to 8 SOL and sold 5 to 15 seconds later, and many were the dev's own wallets. When you copy them, they are already selling to you.

Findings →
Grey area

Chasing the narrative

Buying whatever word is trending. Waves are real but short: "cyber" peaked at 323 new coins in one hour. By the time it is everywhere, the early buyers are leaving.

Narrative radar →
Bad for you

Following calls

Callers are often paid, or bought before they posted. A "profit" screenshot usually shows a gain the poster has not sold yet. Check when the post went up against the chart.

Grey area

Momentum on the curve

Buying coins that fill the bonding curve fast, hoping to sell at graduation. A fast climb on a bundled coin is often insiders pushing it up so they can sell into you.

Grad radar →
Manipulation

Fake profit claims (LARPs)

Giant profit stories and celebrity links spread fast and are rarely proven. A trader's $40M $TRUMP claim was traced on-chain to about $6M (Bubblemaps, Oct 2026). Ask for the wallet.

Level 5 ↑

What devs do

Fair play

An honest launch

No bundle, a small dev buy or none, a dev who sells slowly in the open or holds, and a community that was there before the coin. It exists, it is just rare.

Manipulation

Bundled launch

Insiders buy a big part of the supply in the launch block, then sell into the first real buyers. In our data, 95% of bundled launches were dead within 3 hours, against 87% of the rest.

Level 3 ↑
Manipulation

Serial launching

One wallet launches coin after coin, buys its own supply and dumps it each time. One wallet we caught launched over 300 coins in a day.

Dev check →
Manipulation

Split wallets

The supply is spread over many fresh wallets with 1.5% to 5% each, so the top holders look normal. Bubblemaps shows them joined; check where their SOL came from on Solscan.

Bubblemaps ↗
Manipulation

Volume bots

The same wallets buy and sell over and over to fake activity and get into trending lists. Watch for many trades of almost the same size from a few wallets.

Live tape →
Grey area

Paying for attention

DexScreener profiles, boosts, paid trending spots and paid callers. Not a scam by itself, but it only proves someone paid, not that the coin is safe.

Launch log →
Grey area

Takeovers and relaunches

The dev leaves, and a "community" takes over or launches the same name again. Sometimes real, sometimes the same insiders with a fresh story. Check who holds the new supply.

Grey area

Mayhem mode

A pump.fun mode where an AI agent trades the coin. It is a real feature, but the trade counts do not mean real people are interested. LabRat marks these coins with a caution.

Scam

Fake tools and "verify" bots

Look-alike trading sites, "claim your airdrop" pages and Telegram "verify" bots that ask you to sign something or paste a command. They drain wallets.

Level 7 ↑

Check before you ape.

Paste any pump.fun coin and see who was there first. A few seconds. Free.

Scan a coin →
Link copied