🎓 Memecoins Training

The other side's playbook — how launches, snipers and bundlers actually work, and how to spot every move.

This is the scammers' playbook — handed to you. Not so you can run these plays, but so you recognize every single one and stop being the exit liquidity they farm. None of it is secret: the tools, the mechanics, the numbers are all public. The only reason it keeps working is that most people who ape a launch have no clue what's happening on the other side of their buy. After this, you will.
🔐 Important security notice — read this before you touch a wallet

Most people here don't get rekt by a bad trade — they get drained. The tokens are optional. This part isn't.

  1. "Not your keys, not your wallet" — and that includes non-custodial wallets. A wallet is only truly yours if you generated the seed and nobody else ever saw it. Never fully trust a wallet someone else created, gifted, or "set up" for you — and never park a serious bag in one, no matter how many security audits it "passed." Audits get gamed and keys get skimmed. Keep the real money in a wallet you made yourself; treat every other wallet as borrowed and disposable.
  2. Your private key / seed phrase must NEVER touch the web. Not in a chat, not in an LLM prompt, not in a web search, not in a "wallet checker," not in a support DM — never, for any reason. And don't leave it sitting in your clipboard while you browse: Axiom, GMGN and plenty of others already read your clipboard to grab the token CA you copied — the exact same access lifts a private key you copied too, and so do shady extensions and malware. Be paranoid. Better safe than sorry.
  3. Use a burner wallet for WalletConnect and dApp connections. If you're going to connect to sites and sign things, do it from a throwaway wallet that only holds what you can afford to lose — never your main stash. One malicious approval should cost you lunch money, not your life savings.
  4. Revoke approvals you're not using. Every "approve" you sign is a standing permission for a contract to move your tokens later — it doesn't expire when you close the tab. Review and revoke stale approvals regularly; drains often fire weeks after you forgot you ever connected.
  5. Bookmark the real URLs — don't search, don't click links. Phishing clones buy the top ad and search spots, and fake wallet apps and browser extensions are everywhere. Save the genuine dApp and wallet URLs as bookmarks and use only those — never reach your wallet through an ad, a DM, a "claim your airdrop" link, or a search result.
  6. For real money, use a hardware wallet. The keys stay on the device and can't be exported, so even a fully compromised computer can't move funds without your physical confirmation. Read what you're actually signing on the device screen — not just what the website claims you're signing.
  7. Urgency is the tell — and nobody legit will ever ask for your seed phrase. "Claim now," "you've been selected," "your wallet is at risk," a mod who DMs you first, an APY that's too good — all of that pressure exists to make you skip everything above. Slow down. When in doubt, do nothing.
Chapter 1 · For the newbie

You are the exit liquidity

Read this first. If you take nothing else from the page, take this.

If you're a "normie", the best advice we can offer is to never touch memes. You've probably seen bait posts on X like "this guy turned $5 into $1M — only with memes." They're all false. BS. You will turn into exit liquidity FAST. You'll pick up a new addiction. Please, just stay out.

Solana is the main casino. Even if it was built for a higher purpose, the fee-burning mechanism is exactly the engine of the meme casino — and it runs on the volume memes bring. Roughly 20,000–30,000 new tokens are minted every day. Launching one costs about $4 and thirty seconds, no code, no audit, no liquidity to seed. That frictionlessness isn't a bug — it's the business model.

~98.6%
of pump.fun tokens fit rug-pull or pump-and-dump patterns (Solidus Labs)
~0.6%
of tokens ever "graduate" — the other ~99% die, most within minutes
4.4 min
median time for a "winner" to be decided. The whole game is over before you've read the chart
12 wallets
≈320 launches each drained ~82% of exit-scam liquidity in one 4-month window (Arkham)

Here's the part nobody tells you: the market is engineered so that someone is always holding the bag when insiders sell — and it's built to be you. The dev buys the cheapest supply at the bottom, bots pile in milliseconds later, the chart goes green, retail chases the green candle, and the early wallets dump into that demand. You didn't lose because you "picked wrong." You lost because you were the demand the whole thing was designed to sell into. That is the literal definition of exit liquidity.

Everyone screaming "get in NOW before it sends" needs you to be fast and unquestioning — because speed and hype are the exact triggers the machine is waiting for. If a trade requires you to be first to profit, that first-place slot was already sold to someone with better infrastructure than you'll ever have (Chapter 2). You're not early. You're late by design.

If you're going to play anyway

This site doesn't hold your funds, route your orders, or give financial advice. It shows you what's on-chain so you can decide. Scanning ≠ safe. There is no safe.

Chapter 2 · The machine

How they snipe — and why you're always last

You're not competing against other humans. You're competing against machines and the dev.

A "sniper" is a bot that buys a brand-new token in the same block it's created — before a human can even see it exists. Solana blocks are ~400 milliseconds. You trade in seconds; bots trade in milliseconds. That gap is structural. A fast human cannot close it.

The pipeline that beats you

The biggest sniper is usually the dev. They know the exact mint address and the exact launch moment — an edge no outsider has. They weaponize it to buy their own token in block 0, then sell it back to everyone who shows up after. Chapter 3 is how they do it at scale.

The tools regular traders use

These are real, no-code products. The point isn't to endorse them — it's that even with them, you're still slower than the dev and the pro snipers. Web terminals: Axiom, Photon, BullX, GMGN, Padre. Telegram bots: Trojan, BONKbot, Maestro, Banana Gun, Bloom (some even snipe off X keywords). They make you fast relative to other retail — not relative to the machine that already transacted in block 0.

What actually happens in the first minute

15,000+
tokens sniped in the launch block itself in one month (on-chain study, 2025)
~87%
sniper success rate. The machines rarely miss
85%
of snipes dumped within 5 minutes — often onto the retail that just arrived

The takeaway: block 0 is closed to you — no gRPC feed, no pre-signed tx, no tip auction, no co-location. Every price you can actually get is a price an insider already sold into or is about to. The winning move isn't to play faster; it's to not chase. Patience and verification are the only edge retail has.

Sources: Banana Gun, QuickNode (Jito), Jito docs, pump.fun sniping study.

Chapter 3 · The setup — and the catch

How they launch & bundle — and how to catch them

This is where you turn the tables. Every trick here leaves a fingerprint on-chain.

The launch, in plain terms

On pump.fun, anyone fills a web form (name, ticker, image) and mints a token for ~0.02 SOL in under a minute. There's no seeded liquidity — the token trades against a bonding curve, an automatic formula where price only rises as SOL flows in. The earliest buyer gets the cheapest tokens; every later buyer pays more. That's the structural reason insiders profit off you: they were first, on purpose.

If a token's market cap reaches ~$69k (~85 SOL of net buys), it "graduates" — the curve stops and liquidity migrates to a real DEX (PumpSwap), with the pool's LP tokens burned so migration liquidity can't be pulled. Sounds safe? Two catches: ~99% of tokens never graduate, and the LP burn does nothing about the real theft — the insider dump on the curve, before graduation. Rival launchpads (LetsBonk, Raydium LaunchLab, Bags, Heaven, Believe, Moonshot) mostly just fight over who keeps the fees. The risk to you is identical.

Bundling — faking a crowd

A bundle packages the token creation plus buys from many fresh wallets into a single Jito bundle, so they all land atomically in block 0. The result: one entity secretly holds a big chunk of supply — bought at the cheapest price — but the holder list looks like dozens of independent wallets each holding 1–5%. It's a fake crowd. One person, fifty wallets.

Why they do it: control supply, manufacture a "healthy" distribution, spike early volume/hype, then dump on retail. Variants: bundled snipes, "insider/rat" wallets funded from one source right before launch, and splitting the dev bag across sub-wallets to dodge a lazy top-holder check. Bundler services (Smithii, SolBundler, PandaTool and open-source scripts — some literally advertising "bubble-map-bypassing" modes) make it point-and-click. Bundled launches routinely lock up 30–40% of supply at the bottom. Note: the same-block mechanic has a legit anti-sniper use too — so bundling isn't automatically a scam, but a high bundle % is a serious red flag. The only check that measures it directly is a proper block-0 reconstruction — see how to spot it below. memecoins.expert runs that reconstruction, and every token page here shows its result.

🕵️ How to catch it — the red flags

Every one of these leaves a trace. Run this checklist before you touch anything:

🎁 The initial-bundle reconstruction measures it exactly — the only metric that answers "is it dev-bundled?"

Every other check here is circumstantial. This one is the actual number. A proper reconstruction walks the token back to block 0, rebuilds the launch cohort, and gives you what nothing else does directly: how much of the supply was sniped in the first block (plus the next couple), which wallets did it, and — the honest part most tools skip — how much they still hold vs already dumped. A high bundle % that's still held is one entity sitting on the exit above your head; a high % that's already dumped is the rug that already happened. memecoins.expert runs it; every token page here shows the result under 🎁 Launch bundle.

But a bundle is not a guaranteed rug. Roughly 99% of the time a heavily-bundled launch ends the same way — the hidden cohort dumps on the crowd it manufactured. Every so often, though, a serious project bundles block 0 on purpose: to lock supply away from snipers, or to hold a treasury/team allocation from the start. So read the bundle % as heavily-weighted evidence, not a verdict — a big one still means the launch was controlled by someone, and buying it is a bet that whoever controls that supply won't sell it into you. Usually they do. Size accordingly.

And "no bundle figure" is not "no bundle". Some launches can't be walked back cheaply. When you see nothing, that means nobody measured it — which is a completely different statement from a measured zero, and the one place people fool themselves.

🫧 The holder map catches the fake crowd

A bubble map of the top holders wires the coordinated ones together — shared funder, co-buys, one entity in many costumes — as a connected, coloured cluster against the neutral crowd. That "distributed" holder list stops fooling you the second you see the bubbles joined up. memecoins.expert draws it; Bubblemaps and InsightX do too.

🪤 The shitlist tells you whose wallets those are

This is the part a generic scanner can't do: a bundle % tells you supply was controlled, but not by whom. Every deployer, bundler and top holder can be checked against this list — a public, human-curated record of the wallets and KOLs that have already done it to somebody. A clean-looking launch run by a wallet with three rugs behind it is not a clean launch.

🍯 The honeypot check tells you if you can get out

Quote a small buy and a sell on Jupiter (read-only — no funds moved). "No sell route" = the classic honeypot; a low round-trip % = a hidden sell tax. It answers the only question that matters when you're already holding: can I get out? memecoins.expert runs this on every token it scans.

🔗 The connection check unmasks the cluster

Paste any two wallets on a wallet page and we surface every link — direct transfers, a shared funding wallet, shared counterparties, co-buys. It's how you prove that "50 different holders" is really one operator moving between costumes.

External tools worth knowing too — Bubblemaps, InsightX, Trench (bundle %), RugCheck, and GMGN's bundle/insider/rug ratios. Use more than one. The bundlers evolve to beat detection, so treat every flag as weighted evidence, not a single pass/fail.

Sources: Bubblemaps, Flintr, Solidus Labs, RugCheck.

Chapter 4 · The blind spot

When a clean scan lies

Some rugs pass every check — including ours — and still bleed you to zero. Here's the honest reason.

By now you can spot the obvious traps. But the smartest operators don't trip a single one. Their token has a clean bubble map, LP burned, authorities revoked, holders that look nicely spread out — and it still grinds to zero and takes you with it. That's not the scanner being wrong. It's the scanner doing exactly what it can do, which is less than people assume.

A scanner is a snapshot, not a live camera. Every mainstream tool — Bubblemaps, RugCheck, GMGN, memecoins.expert — answers your question by reconstructing a picture from what's already on-chain and already indexed the moment you ask. It is not watching every wallet, every block, forever. That leaves two blind spots a patient scammer lives inside.

Blind spot #1 — The brand-new launch (the picture isn't there yet)

A bubble map draws a line between two wallets only if there's a transaction already recorded between them. At a launch, the connections are forming block by block, in ~400-millisecond slots, from fresh wallets with no prior history for the map to draw anything from. Free and public RPCs are rate-limited and their indexes run minutes behind the chain — so seconds after a launch, the graph simply hasn't been assembled yet.

Catching every same-block bundle and funding link in real time is a completely different sport: you'd need a dedicated, co-located node streaming every transaction live (Geyser / Yellowstone gRPC, Jito ShredStream) and clustering it per block as it happens — infrastructure that runs hundreds to thousands of dollars a month and that no free scanner runs per token, for everyone. So a green light on a token that's minutes old isn't proof it's clean — it's proof the tools haven't caught up. Re-check it once it's aged; the graph tightens as history piles up and finally gets indexed.

Blind spot #2 — The slow rug (the picture got laundered)

The other move is to let the picture go clean on purpose. Instead of one dump that craters the chart and lights up every scanner, insiders bleed their bag out slowly — hundreds of small sells over hours or days, fed into real buy pressure, so no single candle ever looks like a dump. Then they break the trail: route the proceeds through fresh throwaway wallets, add time delays, and — the killer — round-trip through a centralized exchange. Deposit to Binance, withdraw to a brand-new wallet, and on-chain the connection just ends. The two wallets show zero link.

So by the time you pull up the bubble map, the original bundle has already redistributed into new, innocent-looking wallets, the funding lines are severed, and the map genuinely reads clean — because it only shows the current state, and the coordination has been smeared across time and addresses it can't see. Meanwhile the chart "died naturally" on organic-looking volume. That's a slow rug, and it's built to survive exactly the checks you just learned in Chapter 3.

What you can still watch — over time, not in one snapshot

A slow rug is caught by watching the flow, not by reading a single-moment scan:

The honest bottom line: a clean scan is necessary, not sufficient. It rules out the dumb, instant traps — and says almost nothing about a controlled, laundered exit already in motion. That's true of every tool on earth, ours included: they read a moment, not the future. So treat "brand new and spotless" with more suspicion, not less. Watch the flow over days, not seconds. And size every meme like it's going to zero — because the ones engineered to pass every check are the ones that mean it.

Sources: Solidus Labs, CoinDesk (ZREAL), Bubblemaps, Chainstack (Geyser/gRPC).

Chapter 5 · Your own edge

Run your own sniper — without getting rekt by the tool

Everything above is why the game is tilted. This chapter is the one piece of the playbook that's actually yours to use: a sniper is just fast, automated trading — not a scam. It won't make you the dev (you already read why you can't be in block 0), but it does close the gap with other retail: instant execution, automatic take-profit and stop-loss, hard filters, and no fat-fingered buy at the top. Used with discipline that's a genuine edge. Used as a money printer, it's just a faster way to donate.

200+
bots fighting over one pump.fun launch in the first 500 ms
10–15%
of launch snipes actually land profitably — even for the pros
~1%
per-trade fee on every major bot (Trojan, BONKbot, Photon…)
Read this before you touch anything — this is where people actually lose the bag. The tool is a bigger risk than the trade.
  • Telegram is full of clone bots. Search "Trojan" or "BONKbot" in Telegram and half the hits are impersonators that drain your wallet the second you fund it. Only ever open a bot from its official verified X / website — never from in-app search or a random reply.
  • "Free sniper bot" GitHub repos are frequently wallet drainers. They lift your private key straight out of the .env file and ship it to the attacker — padded with fake stars and forks to look legit (SlowMist documented exactly this on a "solana-pumpfun-bot"). Read the code. Never run one blind.
  • Burner wallet, always. Bot or your own code — fund a fresh, disposable wallet with only what you can afford to lose. Never your main wallet, never your seed phrase, never a key that touches real money.

Track A — No-code, up in five minutes

The fastest way in. Every popular bot works the same way: it makes you a non-custodial wallet (you hold the keys, you sign each trade), you fund it, set your rules, and snipe.

Pick your tool — each tab is the full setup, from zero to your first snipe. Every one of them is non-custodial (you hold the keys), charges ~1% a trade, and — re-read the warning above — must be opened only from its official link.

  1. Step 1 · Sign up Open Axiom ↗ and sign up with Google, email, or wallet. It gives you an Axiom trading wallet — save the recovery key offline and never share it.
  2. Step 2 · Deposit SOL Click Deposit, copy your Axiom wallet address, and send SOL from an exchange or Phantom. Keep ~1–3 SOL if you want to snipe more than one launch without topping up.
  3. Step 3 · Set defaults Set slippage (10–25% for fresh launches) and a priority fee, then turn on Sniper Mode / Buy-on-Migration plus a limit or Sell-on-Migration order so your exits are automatic.
  4. Step 4 · Snipe from Pulse Open the Pulse tab — every new launch in real time with the deployer's history, liquidity and socials. Pick one, enter your SOL amount, confirm.
  5. Step 5 · Scan first Before you confirm, paste the CA into memecoins.expert — bundle %, honeypot, launch cohort — and check it against the shitlist.
  1. Step 1 · Open GMGN Open GMGN ↗ and connect a wallet or use its built-in one. GMGN's edge is on-page bundle %, dev history and top-holder analysis for every token.
  2. Step 2 · Fund it Deposit SOL to your GMGN wallet (or the connected one). Use a fresh wallet, never your main.
  3. Step 3 · Set defaults Slippage (10–25% for launches) and priority, then turn on auto-buy / sniper with a take-profit & stop-loss.
  4. Step 4 · Snipe New / Trending Watch the New Creations feed, read the built-in bundle & holder checks, enter your SOL, buy.
  5. Step 5 · Cross-check GMGN's checks are solid; memecoins.expert adds the launch cohort it doesn't have, and this list tells you whether the wallets behind it are already known.
  1. Step 1 · Open Photon Open Photon ↗ and connect Phantom or Solflare. Photon's edge is the fastest real-time data (HyperSpeed) — the difference on a hot launch.
  2. Step 2 · Connect a burner Photon trades from the wallet you connect, so connect a dedicated Phantom holding only what you'll risk — never your main.
  3. Step 3 · Set presets Pick a fast buy preset, set slippage (10–25%) and priority, and a sell / stop-loss.
  4. Step 4 · Snipe from Memescope The Memescope feed streams new pairs the instant they list. Click one, enter SOL, buy.
  5. Step 5 · Scan first Paste the CA into memecoins.expert before you confirm, and check the deployer against the shitlist.
  1. Step 1 · Open the bot Open the official bot — @solana_trojanbot ↗ — and tap Start. Use only this link; clones copy the name. If in doubt, get it from Trojan's site (trojanonsolana.com).
  2. Step 2 · Get your wallet It creates a Solana wallet for you. Back it up, never share the seed, and enable 2FA under Settings.
  3. Step 3 · Fund it Copy the wallet address and send a small, disposable amount of SOL.
  4. Step 4 · Buy / Sell settings Settings → Buy Settings (preset SOL sizes, slippage 10–25%, priority fee) and Sell Settings (take-profit / stop-loss).
  5. Step 5 · Snipe Paste a token CA to buy now, or turn on Sniper mode with your filters — after you've run the CA through memecoins.expert.
  1. Step 1 · Open the bot Open the official bot — @bonkbot_bot ↗ (also linked from bonkbot.io) — and tap Start. It builds you a fresh SPL wallet.
  2. Step 2 · Back up the key Settings → export your private key and store it offline. It's non-custodial — only you hold it, so nobody can recover it for you.
  3. Step 3 · Deposit Copy the address and send 0.1–0.5 SOL to start.
  4. Step 4 · Buy size + slippage Settings → a default buy (e.g. 0.1 ◎) and slippage 10–15% for new launches.
  5. Step 5 · Buy Paste a CA (or a pump.fun / DexScreener link) → tap Buy; it routes via Jupiter for the best price. Scan the CA with us first.

Every link above is the official one, carrying a referral where a program exists — it costs you nothing and helps keep this free. Always double-check you're on the real site / bot before you fund a wallet.

Track B — Build your own (for devs)

Want to actually understand and tune what's happening? Run your own. Don't start from a random repo — start from a maintained one published by a real infrastructure company:

  1. Step 1 · Trustworthy base chainstacklabs/pumpfun-bonkfun-bot ↗ — Python, open-source, maintained by Chainstack (kept current with pump.fun's 2026 program change), talking to the chain directly with no third-party trading API in the middle. Its own README warns that its Issues tab is full of key-stealing scam forks — trust the source, not lookalikes.
  2. Step 2 · Install Clone it, install uv, run uv sync, then copy .env.example → .env and drop in your RPC endpoints and a burner private key.
  3. Step 3 · Configure a strategy Edit a .yaml template in the bots/ folder — which launches to buy, buy size, take-profit / stop-loss, max RPS — then run pump_bot.
  4. Step 4 · Read every line first Before it ever sees a funded key, read the code end-to-end and confirm nothing phones home. When in doubt, run it against a burner holding pocket change.

The real edge isn't the bot — it's the RPC

This is the part beginners skip and pros obsess over. A public RPC will not work: on a hot launch your transaction is slow or silently dropped and you lose to everyone on better infrastructure. The bot is a thin wrapper — the node underneath wins or loses the block. Roughly in order of what matters:

The honest expectation. None of this makes you the dev. They minted the token, they were in block 0, and on a bundled launch they're already sitting on the cheap supply you're buying. A retail sniper makes you faster and more disciplined than other retail — that's it, and it's still worth having. Keep buys small, let the stop-loss do its job, assume most snipes lose, and never point it at a wallet you can't afford to zero. The people printing on launches aren't lucky — they run better infrastructure and better risk management than you. Now you know what both look like.

Sources: Dysnix (sniper stack), RPC Fast, QuickNode, Chainstack bot, Helius Sender, SlowMist (drainer bots).

Chapter 6 · The truce

Safer plays — how to lose less

Everything above is why the game is rigged. This chapter is the truce: if you're going to play anyway, here's how to not torch your whole stack on the first red candle. None of it makes you win — it makes you last longer, and in a casino, lasting longer is the only edge that compounds.

New pairs? Small size. Always.

Never put real money on a fresh pair. You already read Chapter 4 — on a brand-new token the scan can't even tell you the truth yet, so you're flipping a coin that's weighted against you. Treat every new-pair buy as money that's already gone, and size so that "gone" doesn't hurt.

Take profit, keep a moon bag

This one move separates the survivors from the exit liquidity: take your money out, let house money ride. Turn 0.2 SOL into 0.4? Pull your original 0.2 (or more) back into your wallet, and let the rest run as a free "moon bag." Now you literally cannot lose on that trade — worst case the moon bag zeroes and you broke even; best case it 10×'s on money that was never at risk.

And don't be greedy. The chart that's up 3× doesn't owe you 10×. Greed will kill your portfolio faster than a rugger can hit "remove liquidity" — and he's got that on a hotkey.

Trading real size? Only established tokens.

If you're deploying serious money, don't touch fresh mints — go where there's actual liquidity, real history, and a settled holder base. Then do the homework:

✅ On established tokens, the scanners actually work

The blind spots from Chapter 4 are a new-token problem. On an established token the picture has fully developed — full holder history, distribution has settled, wallet connections are traceable, and RugCheck and memecoins.expert return accurate reads. So on the exact tokens where you're risking real size, the tools are reliable. There's no excuse not to run every check you can find — stack them, use whatever tool you can get your hands on.

The house rules that keep you in the game

🧢 NFA & other legally required noises. This is a website that keeps a public list of degenerates, written by degenerates, called SolanaShitList.fun. Nothing here is financial advice — it is barely advice, and legally it's closer to graffiti. We are not your financial advisor, your fund manager, your dad, or your therapist (though we'd charge less than all four). If you make a trade because of anything on this page, that's a decision you made with your own two hands, and the candle owes you nothing. Do your own research, size like an adult, and if it all goes to zero — and it might — that's between you, your wallet, and the God of Liquidity. You have been warned, probably too late. gm. 🫡
The whole edge they have is you not knowing this.
Now you don't have that excuse. Before you ape anything — paste the token and see who's really holding it.
🔍 Scan a token

← Back to the leaderboard