Solana forensic intelligence

Spot the rug before it's pulled.

RugHawk reads the launch block, traces the deployer and audits every authority in seconds. Then Cloudflare Workers AI tells you what it means.

Try:BONK PYUSD (Token-2022)Jupiter program

Cloudflare Workers AI · AI Gateway · Vectorize · Browser Run · QuickNode

Live from the hawk

Latest tokens flagged HIGH or CRITICAL

What the hawk sees

Scanners show you a score. RugHawk shows you the operator.

Most rug checkers stop at authorities and holder percentages. The money is lost in the first minutes of a launch, so that is where RugHawk looks hardest.

Launch forensics

Reads the creation block itself

Dev buy at creation, every wallet that bought in the launch block, and how many seconds after launch the deployer started dumping.

Deployer dossier

Who launched it, really

Wallet age at launch, serial-launcher detection, and the funding trail back to an exchange or a chain of burner wallets.

Authority audit

Every power the issuer still holds

Mint and freeze authority, mutable metadata and all Token-2022 traps: permanent delegate, transfer hooks, fee hikes, default-frozen, pausable.

Holder intelligence

Concentration without the noise

Top holders classified as pools, bonding curves, burns, lockers, exchange custody or plain wallets before any percentage is computed.

Scam DNA

Same scam, new mint

Every token identity is embedded with Workers AI and stored in Vectorize, so the 40th copy of a flagged narrative lights up instantly.

Site & logo forensics

What the website really shows

Browser Run renders the token website, Llama 4 Scout reviews it for drainer lures and checks the logo for brand impersonation.

Wallet Shield

Kill drainer approvals. Get your rent back.

Connect a wallet to see every live token approval, every token its issuer can seize, and every empty account holding your SOL deposit. Revoke and reclaim in one signed transaction.

Open Shield

Ask the Hawk

A forensic copilot that cannot be talked into lying.

Paste an address and ask anything. The deterministic engine decides what is dangerous; Workers AI explains it, grounded in AI Search, behind Prompt Guard and Llama Guard.

Ask a question

Field-tested