The pattern that made us angry
Every few weeks, the same Telegram message: “hey is this token safe?” Three hours later: “shit, it just rugged.”
Honest people throwing $200, $2,000, sometimes $20,000 at contracts whose first transaction was a hidden mint or a transfer-to-zero from the deployer. The signals were always there. They were just buried under a fresh logo, a Discord with 800 paid members, and a Telegram raid that looked like organic excitement.
The tools that existed — Etherscan, Token Sniffer, GoPlus — each gave one slice of the picture. Nobody combined the on-chain signals into something a non-developer could read in ten seconds. So we built it.
What RektRadar actually does
You paste an Ethereum contract address. We score it 0–100 in real time across 80+ flags pulled from five independent angles:
- Contract analyzer — bytecode patterns, ownership, fees, blacklist functions, hidden mint, proxy upgradeability.
- Network graph — every wallet that funded the deployer, every other contract that deployer ever shipped, clusters of related rugs.
- Mempool watcher — sandwich bots, rug-pull preparations visible before they confirm on-chain.
- Liquidity & holders — concentration, lock duration, sell tax history.
- Forensics — for premium users, a deeper post-mortem with the funding trail and exit pattern.
The whole pipeline runs on three Ethereum nodes (one mainnet, two mempool shards), 18 microservices, and a Postgres that has indexed every analyzed contract since the day we turned it on.
The path here
Phase 0 — A weekend script
The first version was 200 lines of Python and a notebook. It checked three things on a contract. We sent it to friends in Telegram when they asked. That’s it.
Phase 1 — “Maybe this is a product”
Once we had a few hundred contracts indexed, the signal got real. We could group deployers, see whose addresses were funding new rug factories, and predict the next launches before they even paired on a DEX.
We rewrote the analyzer in TypeScript, split it into services (contract-analyzer, scoring-engine, graph-crawler, mempool-watcher), and put it behind a Vue frontend on app.rektradar.io.
Phase 2 — Distribution
Building a scanner is one problem. Getting it in front of the user at the moment they’re about to ape is a completely different one. That’s where we are now: bots on Telegram and Discord that let you scan a token without leaving your group chat, listings on the major DEX trackers, SEO so a search for “is $TOKEN safe” lands you on a real analysis page instead of an aggregator.
What’s coming
- Per-symbol hub pages (
/scam/<symbol>/<address>) — because a query like “PEPE scam” shouldn’t land you on one arbitrary contract; we list every $PEPE that’s been flagged so you know which one you’re looking at. - Real-time alerts on factories you care about — get a ping the moment a new contract from a known scam deployer hits the mempool, not after the rug.
- Cross-chain — Ethereum first, but the same architecture extends to BSC, Solana, Base. The flags differ; the model doesn’t.
- This blog. 70% case studies (decortiquage of real rugs we caught), 30% infra notes (how the pipeline scales, what we got wrong).
Who we are
A small team that got tired of watching the same scam repeat itself. We’re fully self-funded, run our own infra (three Ethereum nodes, no third-party RPC dependencies for the critical path), and ship what we’d want to use ourselves.
If you find a token, scan it on RektRadar. If you spot a flag we miss or get wrong, tell us. If you want to follow the case studies, the next post is dropping next week.
The principle: the scam was always visible on-chain. We just needed to read the chain instead of the marketing.