Robinhood Chain · every token · every trade
The tape never sleeps.
We follow Robinhood Chain one second behind the chain head, price every trade from the pool it settled in, and keep score of what happened next. Now the whole thing is for rent.
API · safety
Rug Radar: know what a coin is before you touch it
A boosted-tree model scores every young coin on the odds it gets rugged within six hours. Held-out AUC 0.936. One bytecode template alone has been blocked 5,041 times before its coins could reach anyone's feed.
- A rug probability with a plain band, LOW to AVOID, calibrated on held-out data and refit monthly.
- Bytecode template match at mint: factory rugs caught before the first trade, with zero added latency.
- Liquidity custody from the pool registry: locked, burned to the dead address, or pullable by a wallet.
- A sell check that trades through hooked v4 pools, because a plain transfer test lies there.
Weakest where coins are one to seven days old with locked liquidity (AUC 0.70 to 0.80 there). The response says which segment it came from.
How often each band actually rugged
Held-out rug rate within six hours, by the band the model assigned beforehand. 9,074 scored moments.
Show as table
API · market data
The Tape: every trade, one second behind the chain
Our follower rides the node's head feed, decodes v2, v3, v4 and bonding-curve legs, and prices each trade from its own pool. No third-party price in the loop, which is why hooked pools and curves that other listings leave unpriced show up here with a dollar value.
- REST and WebSocket: trades carrying the pool and the leg each price came from, plus 5-minute and hourly candles.
- A pool registry of 976K v4 pools, 333K of them hooked, across 25K distinct hook contracts.
- 72K bonding curves priced natively, graduations included.
- Bulk history back to the first block we followed in May 2026.
Priced trades per day
Swaps the follower priced from their own pool, last 30 days. Today is partial.
Show as table
Report · launchpads
Launchpad League: where the coins that lasted come from
We attribute every coin to the launcher that minted it, then score what happened next. On LONG, 59% of coins reached 5× within a week. On the biggest launchpad it was 11%. That gap is the whole report.
- A monthly scorecard: launches per day, share that reaches 100 trades, 5× rate, blocked rate.
- A registry of 37 named launchpads with their contracts, pool mechanics and where the LP went.
- Hook-by-hook breakdown for v4 launchers, so a fee-skimming hook is visible before you buy.
Share of coins that hit 5× within 7 days
Final outcomes only. Launchpads with at least 100 scored coins.
Show as table
Who is launching
Coins on file, coins with 100+ trades, and launches in the last 30 days.
API · wallets
Smart Money: 84K wallets, graded on what they bought early
Every wallet that bought a coin in its first 30 minutes gets a record: how many of those coins went on to 10× or 20×. Wallets that fund each other fold into one entity, so an operator running 665 wallets reads as one operator.
- A wallet profile: early hit rate, realized PnL, which entity it belongs to.
- Who bought in the first 30 minutes, for any coin, with what they did next.
- A live feed of named wallets' buys and exits, delivered to Telegram or your webhook.
Feed · signals
Signals, with receipts
Seventeen detectors watch the tape for coins waking up, momentum, smart-money entries and exits. Every delivered card is scored at 1, 6 and 24 hours and the ledger is public. About a third of cards hit 1.5× within six hours. The median card closes below entry. We would rather you knew.
- Real-time cards with the model's read attached, so you can filter by what has worked.
- The outcome ledger as data: entry price, highs and lows at 1, 6 and 24 hours, for every card ever sent.
- Charts drawn from our own tape under every card.
The last 120 cards we sent
Each dot is one delivered card, coloured by where it closed 24 hours after entry. Hover for the numbers.
The whole ledger, by tier
What's in the box
One chain, counted properly
Everything below comes from our own follower on a Robinhood Chain node. No scraped listings, no borrowed prices.
Counts from the snapshot taken today. The ledger starts May 2, 2026.
Pricing
Pick a size. Cancel whenever.
Prices are our opening plan. Early-access members lock theirs in for a year.
- Launchpad League, every month
- Signals feed, 24 hours delayed
- Daily chain stats
- 100 API calls a day
- Signals in real time, with the read attached
- Smart Money lookups, 500 a day
- Rug Radar checks, 500 a day
- Telegram delivery
- Everything in Pro
- REST and WebSocket keys, 50K calls a day
- Rug Radar API with custody facts
- Candles and pool registry in bulk
- Outcome ledger as data
- Parquet drops of the full tape
- Read-only SQL on a replica
- Custom datasets and backfills
- SLA and a human on call
Early access
Get on the list
We are letting people in a few at a time while the serving layer grows up. Tell us what you would use and we will put you in the right batch.
Questions
The small print, in a normal font
Where does the data come from?
From our own follower on a Robinhood Chain node. It reads every ERC-20 transfer and swap event, decodes the pool each trade settled in, and prices the trade from that pool. Metadata that coins publish about themselves is harvested on chain too. Nothing is scraped from listing sites.
How fresh is it?
The follower runs about a second behind the chain head. Candles close on their own clock. The counts on this page are a snapshot, regenerated on a schedule, and say when they were taken.
Which chains?
Robinhood Chain, end to end. We have followed other chains before and keep their cursors, so a second chain is a switch, not a rebuild. We will add one when the first product is paying for itself.
How clean is it?
Honestly: a data-quality engine runs hourly, logs every finding, and fixes what it can reversibly. Every price carries the pool and event it came from so you can audit it. The transfer ledger misses a small share of logs, which is why holder counts are verified on chain before we sell them. We filter the handful of absurd dollar values that any on-chain tape produces, and we show you the filter.
Is any of this investment advice?
No. The signals are a screening feed with a published hit rate that is far from perfect. The stock-token prices are what people paid on chain, not an official quote. Nothing here tells you what to buy.