Protecting the Robinhood Chain community

Find out if a token is a honeypot before you buy it.

HoneyCanary is an AI agent that checks tokens on Robinhood Chain for honeypots and rugs. Paste a token's contract address and get a low, medium or high risk rating, with the reasons in plain English, and a report card you can share with your friends before they buy.

Built to help and protect the community: checks are free, because the agent burns its own CREDIT, the tokens that pay for its AI inference, on every scan. Every result is public for everyone.

How it works Why we built it

Check a token

Coming soon

The checker opens here once the agent's API is deployed.

What you get back

Low riskNothing stood out. A test sell went through and the scan found no warning signs.
Medium riskSomething to watch, such as code that can still be upgraded, liquidity that wallets can pull, or a few wallets holding much of the supply.
High riskSelling is blocked or loses half or more of what you paid, or the liquidity is gone.

A shareable report card. Every result comes as an image you can send to friends before they accidentally buy a honeypot.

How it works

One scan per token, shared by everyone.

The agent runs the scan and publishes what it finds, so one check protects the whole community: the next person to look at the same token gets the answer straight away.

01
Paste an address
Paste a token's contract address on Robinhood Chain, the Arbitrum-based L2 with chain ID 4663.
02
The agent scans it
It simulates a buy and a sell, reads the pool, looks up the deployer and the launch, and inspects the contract.
03
You get a report card
Low, medium or high risk, with the reasons in plain English, plus a report card image you can send to friends before they accidentally buy a honeypot.
04
Everyone gets the same answer
Results are public and shared, so a token is scanned once for everyone, not once per person. A result more than a day old is rescanned on the next check, because liquidity and owner settings change after launch.

What a scan checks

It tries to sell before you do.

Every scan reads the token as it is on the chain right now: its pool, its deployer, its launch and its code.

Can you actually sell?

In one read-only call, a simulator contract buys the token from its deepest ETH or WETH pool on Uniswap V2, PancakeSwap V2, Uniswap V3 or Uniswap v4, sends some to a fresh wallet, and sells the rest back. It measures the buy, sell and transfer tax and the round-trip loss, which catches Uniswap v4 hooks that skim swaps without any token tax.

Nothing is deployed and no funds move: the simulator is injected into that one call. Its addresses are randomized on every scan, so a token can't whitelist it to pass the test.

Liquidity

Whether the pool's liquidity has been pulled, and whether wallets hold the LP tokens and can still pull it.

Deployer

Who created the token, whether it came from a brand-new wallet, and how many other tokens that wallet has created.

Bundled launches

The wallets that got the token in its first seconds, and how much of the supply they and the deployer still hold.

Holders

The actual balances of the wallets holding the token, to flag a single wallet or a small group holding much of the supply.

Contract

Privileged functions such as mint, pause, blacklist and fee setters, including inside proxies and minimal clones, and whether the code can still be upgraded.

Verdict

A model writes the explanation, while deterministic rules set a floor under the rating. Neither the model nor text planted in a scam contract can talk a rating down.

Built on Orbio

An agent that pays its own way.

HoneyCanary runs on Orbio. It explains its verdicts with a model on Orbio's inference gateway, pays for its own scans in CREDIT, and reads the chain through Orbio's tools.

Onchain inference

The explanation in each verdict comes from a model served through Orbio's inference gateway. It sits on top of the deterministic risk rules, which set the floor for the rating.

CREDIT for autonomous spend

The agent has its own wallet and derives its Orbio API key from a wallet signature.
It checks its CREDIT balance before every scan and skips the scan when the balance is too low.
The exact CREDIT each scan cost is recorded next to the public result.
When the balance drops below a floor, it can top itself up by buying CREDIT with USDG.

Agent tooling

Every chain read goes through Orbio's chain.read tool on Robinhood Chain: the buy and sell simulation (eth_call with state overrides), the pool, deployer and launch lookups, and proxy storage reads.

Why we built it

Have you ever been burned by a honeypot?

A lot of people have. You buy a new token, and then you can't sell it, or the liquidity disappears right after you buy. We built HoneyCanary to help and protect the Robinhood Chain community: the agent burns CREDIT on AI inference to check tokens, so people don't burn their own money on a honeypot.

Check before you buy

Some new tokens are built to take buyers' money: you can buy but never sell, the deployer drains the pool right after people buy in, or a hook quietly skims most of every swap. Most people can't read a contract or trace a deployer, so they find out by losing money.

We burn CREDIT so you don't burn money

CREDIT is what the agent spends on Orbio: the tokens that pay for its AI inference and its chain reads. We'd rather the agent burn CREDIT checking a coin than have someone burn their own money finding out the hard way. Every check is paid for by the agent, never by the person asking.

A safer community

Every result is public, so one check protects everyone who looks at that token after. Honest launches stand out, scams get caught before they spread, and people who would have been burned stay in the community.

HoneyCanary gives the Robinhood Chain community a shared, public answer to “is this a honeypot?”

The same answer for:

Buyers
Communities
Launchpads
Wallets
Bots