Lorg Check
Know it's a trap before you buy it.
Lorg Check is the safety check your agent or trading platform runs before every trade. One call. About a second. A clear answer: allow, caution or block, and the reason why.
The problem
Buying is easy. Selling is where the trap is.
Most new tokens are not what they look like. Some let you buy and never let you sell. Some let the creator print unlimited tokens and dump them on you. Some look healthy until one wallet takes all the money out of the pool.
A person might notice. An agent trading on its own at three in the morning won't.
How bad it is
Three in four new launches with real money in them are scams.
We followed 103 new tokens on Base for a week. 77 were scams. One group alone launches about 100 trap tokens a day, and each one is emptied within 20 to 40 minutes of going live.
A check that takes half an hour to notice a new token protects nobody.
What we do
One question. One answer.
Before a trade, your platform sends us the token's address. We check it against live blockchain data and answer in about a second:
Do not proceed. This pool keeps 99.99% of whatever a buyer sells.A real answer about a real token
Your code reads one word. Your user reads one sentence. Ask one yourself:
The four answers
Four answers, and what to do with each
Nothing to stop you.
We found nothing that should stop you.
Trade.
Real risks, listed.
There are real risks and we list them.
Warn the user, or keep the trade small.
A trap.
It is a trap, or a sanctioned address.
Don't trade. Show the reason.
Not verified.
We couldn't finish a check we can't do without.
Treat it as not verified. You are not charged.
How we know
We don't look the token up. We try to sell it.
Most checkers pass on what a list says about a token. We run the trade: a buy and a sell, through the same route a real wallet uses, in a safe copy of the blockchain where no money moves. If the sale fails, or most of the money disappears on the way out, we have seen the trap ourselves.
Then we read everything else that can hurt a buyer: who can change the token's rules, who can withdraw the pool's money, who made it and what they did before.
What we catch
What we catch
- Tokens you can buy and can't sell (honeypots), including the newest kind, hidden in the pool instead of the token, which most checkers can't see at all.
- Sell taxes that quietly keep a large share of every sale.
- Hidden printing presses: a creator who can make unlimited tokens and sell them into the pool.
- Pools one wallet can empty: the money buyers trade against, sitting where one person can take it.
- "Locked" liquidity that isn't: a lock that has already run out.
- Fakes: a token using the name of a well-known one, or passing itself off as USDC, USDT or ETH, even with its name spelled in look-alike letters.
- Owner powers: someone who can freeze your tokens, block your wallet or change the fees after you buy.
- Bad addresses: sanctioned wallets, wallets that forward everything they receive to a thief, and look-alike wallets planted in your history with fake transfers, so that your next payment goes to the copy.
- A trade too big for the pool: tell us the size of the trade and we try it at that size. A sound token can still be the wrong place for a $50,000 order.
The proof
Tested on live traps, next to the tools people already use
We watched Base for new tokens over three sessions and asked every checker about each one within the same few seconds. Each trap was confirmed first with a real wallet that bought, sold, and got nothing back.
Two widely used free checkers flagged 1 between them. In one session, 15 new tokens were emptied within two hours of launch. We had blocked or warned on all 15. One of those checkers had called all 15 fine.
Live traps: three sessions on Base, 4–5 October 2026 · 38 of the 39 are one scam family · see the evidence
Honest answers
An answer you can rely on, because we tell you when we can't give one
- We never guess. If a check we need can't run, the answer is "unknown", never "allow".
- You only pay for answers. An "unknown" or an error costs nothing.
- Every answer is signed. It carries the exact moment it was true and a signature anyone can verify, so you can prove later what we told you.
- A new token isn't a bad token. We don't block something just for being new. When we block, we name the reason.
Who it is for
Built for the platforms that trade for others
Trading bots and agent platforms.
Put one check in front of every buy. Your users stop losing money to traps your current check calls clean, and you get a weekly report of everything we caught for them.
Wallets and agent toolkits.
Give every agent built on you a safety check without building one.
How to start
One call. Nothing to install.
Send the token's address before the trade. Read one word back. Most teams have it working in an afternoon.
Start with a free pilot: four to six weeks, on your real traffic, with a named person to talk to and a report every week of what we caught.
Three ways to use it
The API.
One HTTPS call before the trade: paid per check in USDC with no account, or with an API key.
MCP.
Add Lorg Check as a tool in Claude, Cursor and other MCP clients, so an agent asks before it trades. It needs a key. Try it with the public one.
Inside your platform.
Built into your bot, wallet or agent toolkit, so every one of your users is checked without doing anything.
Check before you trade.
Base and Ethereum today. More chains to come.