- $SpongeBobinstant+0%
- $SharkTankinstant+4%
- $HERMESinstant+17%
Your trenches, in one tab
The feed, coin pages with charts, your watchlist and the exit coach. Add it to your home screen.
QuantBase checks every pump.fun coin for the signs traders look for (bundled launch, serial dev, copycat name) and shows how coins like it really did. Then the exit coach tells you when to take your initials, because 71% of coins that doubled were back under entry six hours later1.
7 days free, then $19 a month · no fee per trade · cancel any time
The curve was bought out within seconds of creation. That is the creator's own money, not demand.
The curve was bought out within seconds of creation. That is the creator's own money, not demand.
The curve was bought out within seconds of creation. That is the creator's own money, not demand.
The curve was bought out within seconds of creation. That is the creator's own money, not demand.
The curve was bought out within seconds of creation. That is the creator's own money, not demand.
The curve was bought out within seconds of creation. That is the creator's own money, not demand.
The curve filled from outside buyers. That is the rarer kind of graduation.
The curve was bought out within seconds of creation. That is the creator's own money, not demand.
The curve was bought out within seconds of creation. That is the creator's own money, not demand.
The curve was bought out within seconds of creation. That is the creator's own money, not demand.
The curve was bought out within seconds of creation. That is the creator's own money, not demand.
The curve was bought out within seconds of creation. That is the creator's own money, not demand.
Every course and call group repeats the same rules. We measured each one on the coins that graduated, buying two minutes after graduation and checking an hour later. Some hold up. Some are backwards.
| What they say | What we measured | n | Up after 1 h | Verdict |
|---|---|---|---|---|
| Every graduate | baseline | 512 | 14.8% | |
| Avoid copycats (“vamps”) | Name used 3+ times this week, with a Twitter link | 94 | 7.4% | Holds up |
| Avoid serial devs | Creator launched 10+ coins before | 117 | 10.3% | Holds up |
| Avoid a big dev buy | Creator's own first buy of 3 SOL or more | 164 | 12.2% | Holds up |
| Follow devs who graduated before | Creator had a graduate already | 106 | 9.4% | Backwards |
| A Twitter link means it’s legit | Coin links a Twitter account | 189 | 9.5% | Backwards |
| Only trade hot markets | Launch rate in the busiest hours | 650 | 11.4% | No effect |
| Original name, first-time dev, no Twitter link | All three together | 99 | 22.2% | Best we found |
Even the cleanest coins were up an hour later only about one time in five. The red flags tell you what to skip. The exit is where you keep a win.
Measured 2 Oct 2026 on 512 graduates over 14 days; the other half of the data was held back to check results before we use them. Past rates may not repeat. Nothing here is a buy signal.
Open a coin, read what the creator did, tell the coach what you bought, and let it watch the exit for you.
The chart won't show you that the dev bought out their own launch in a tenth of a second, has rugged twenty coins before, or that thirty other coins used the same name this week. QuantBase does, on every coin.
A curve bought out in seconds is the dev's own money. 99% of those were under entry six hours later2.
Every earlier launch by the same wallet, and by wallets funded from the same source: 182,134 devs on record.
How many coins reused the name, the ticker and the image in the last seven days.
The free Lens extension shows the check beside the coin on Axiom, pump.fun, DexScreener, and on contract addresses posted in X and Telegram Web. Get Lens.
Measured on every launch we recorded, losers included. Each figure carries its sample size.
of coins that doubled within an hour were under the entry price six hours later.
n = 266 · measured 26 Sep 2026
of launches reach a real trading pool at all.
n = 658,604 launches
of organic graduates were up one hour after graduation.
last 14 days
were still up a day later.
last 14 days
of coins that doubled were still at 2x or more a day later.
measured 26 Sep 2026
The win is rarely in the pick. It is in leaving on time.
Every trader on YouTube says it, and almost nobody does it: sell half at 2x, trail the rest, cut the losers. Set your rules once. The coach watches every coin you hold and alerts you the moment a rule fires. You still press sell.
Both charts are drawn examples of the rules, not recorded coins. Fees and slippage are not included.
Every coin our scout would have flagged is scored by what happened next. Most lost. That is why the feed blocks whole categories, and why we don't sell picks.
| Coins, last 14 days | n | Hit 2x | Under entry at 6 h | Median at 6 h |
|---|---|---|---|---|
| Curve filled by outside buyers | 66 | 10.6% | 87.9% | 0.038× |
| Creator bought the curve | 3,967 | 6.4% | 99% | 0.005× |
| Worth $150–300K at entry | 535 | 10.3% | 94.6% | 0.010× |
| Worth $300–600K at entry | 3,409 | 6% | 99.5% | 0.005× |
| Worth $600K+ at entry | 89 | 1.1% | 96.6% | 0.003× |
| Every coin the scout flagged | 4,033 | 6.5% | 98.8% | 0.005× |
Priced two minutes after graduation and again at one and six hours, on the same pool. Snapshots miss peaks between them.
Of the 52 trading videos we studied, almost none showed the losing trades, and nearly all were selling a referral link or a paid group. We charge a flat price and earn nothing when you trade, so telling you to skip a coin costs us nothing.
Terminal fees are as published in their own documentation, September 2026; they vary by product.
The feed, coin pages with charts, your watchlist and the exit coach. Add it to your home screen.
Which rule, on which coin, and the numbers behind it, in your alert inbox.
The same evidence over the Model Context Protocol, so an assistant answers "what happened to coins like this one?" with measured numbers, not vibes.
# ask from any MCP client dex_coin(mint="YWLbxvTg…pump") # answer, shortened kind: instant coins like this, last 14 days hit 2x: 6.5% under entry at 6 h: 98.8% n: 4,033Connect an agent
Open any coin and the launch evidence comes before the chart. Then tell the coach what you bought, and it watches the exit.
And on pump.fun, DexScreener, and contract addresses posted in X and Telegram Web.
Get LensNo buy calls, no price targets, no "gems". The data says nobody has them.
We never hold your coins or your keys. Sign-in with a wallet only signs a message.
The coach tells you a rule fired. Selling stays your decision.
Rules limit losses and keep part of a win. They do not make a bad coin good.
Check coins for red flags. No card.
The live feed, the full odds on every coin, and an exit coach that tells you when to take profit.
A separate product for testing trading strategies on exchange data.
Prices are in US dollars. Your card may be charged in your local currency, plus tax where it applies. Cancel any time; access runs to the end of the paid month.
An exit bot that applies your rules from your own wallet. It is not on sale, and it won't be until the evidence product has earned it.
One email when it opens. No card, nothing else.
No. We never tell you what to buy. QuantBase shows the red flags on any coin you are looking at, the odds for coins like it, and when your own exit rules fire. The picking, and the selling, stay yours.
Yes. The Lens extension shows the check beside the coin on Axiom, pump.fun and DexScreener, and on contract addresses posted in X and Telegram Web. The console works on any phone.
We record every pump.fun launch as it happens, then price each graduated coin at fixed times on the pool it graduated to. 658,604 launches and 2,522,126 price checks since 2026-09-20.
A pump.fun coin starts on a bonding curve. When the curve fills, the coin moves to a real trading pool. About 2.8 in 100 launches get that far.
We don't claim that. Most meme coins lose, and so do most of the coins in our own scorecard. The product shows you the odds and helps you keep part of a win.
No. The product is read-only. If you sign in with a wallet, it signs a message to prove the address is yours; it cannot move funds.
The record starts on 2026-09-20 and covers one launchpad. Prices are snapshots, so peaks between them are missed. Past rates may not repeat.
From the billing page in the console. Access runs to the end of the month you paid for.
Free to start. Paste any contract address and see the red flags in seconds.
Check a coin free1 188 of 266 pump.fun graduates that doubled within one hour, priced on the same pool; measured 26 Sep 2026.
2 n = 3,964 graduates whose curve filled within seconds of creation, last 14 days. The median was 0.005× of entry at six hours.