Documentation

How HOODRADAR turns a contract address into a number you can act on — and why that number means the same thing to everyone who reads it.

01 Overview

HOODRADAR watches every token on Robinhood Chain, scores each one from 0 to 100, and pushes the result to your terminal and Telegram. The score is produced by a pure function: no network calls, no clock, no randomness, and no AI in the number itself.

That property is the whole point. The same inputs always produce the same score, so any historical call can be recomputed and checked. A dispute becomes a replay rather than an argument.

One projection, one number. The radar feed and the scorecard both route through the exact same scoring path. A token cannot show one score on the list and a different score on its own page.

02 The Radar Score

Every token is scored across three dimensions, weighted to 100 points:

  • Security — 40 points. Is the contract safe to hold and sell?
  • Distribution — 30 points. Is the supply spread, or concentrated in a few wallets?
  • Momentum — 30 points. Is real activity building, or is the chart a mirage?

Sub-scores degrade smoothly rather than snapping between full marks and zero. A token at 22% top-10 concentration is meaningfully different from one at 39%, and a cliff would erase that difference.

AI writes the words, never the number. A narrative model explains a finished scorecard in plain English. It runs strictly after scoring and cannot move a single point. A model that could change a score would make every score irreproducible, and an irreproducible score cannot be audited.

03 The security gate

Before any points are counted, four conditions void a token to a score of 0 and a GHOST verdict. There is no partial credit and no chart that overrides it.

ConditionResult
Honeypot — you cannot sellScore 0 · HONEYPOT
Contract blocks sellingScore 0 · UNSELLABLE
Sell tax above 25%Score 0 · PUNITIVE_SELL_TAX
Security status unknownScore 0 · SECURITY_UNVERIFIED

The last one matters most. When the data provider has never simulated a sell, it returns an unknown, not a clearance. HOODRADAR treats unknown as unknown and voids the token. Absence of a warning is not a guarantee of safety. This is the one rule in the system designed to protect money rather than accuracy.

04 The dimensions

Security · 40

  • Contract source verified — +10
  • Ownership renounced — +10
  • Liquidity locked or burned, saturating at 90% — +12
  • Zero round-trip buy/sell tax, decaying to nothing at 10% — +8

A burned liquidity pool counts in full. A burn is stronger than a timelock, because a timelock eventually expires and a burn never returns.

Distribution · 30

  • Top-10 holders under 15%, zero at 40% — +12
  • Holder count on a log scale, ~7.5 at 1,000 holders — +10
  • No single wallet above 5% — +8

Liquidity pools are excluded from the wallet check. A pool routinely holds more than any human, and counting it as a whale would penalise exactly the tokens whose liquidity is deepest.

Momentum · 30

  • Current-hour volume against the trailing 24h mean, saturating at 4× — +10
  • Holder growth, normalized to a 15-minute window — +10
  • Liquidity depth against market cap, saturating at 20% — +10
Holder growth is unmeasured on first sight. A token seen for the first time has no prior snapshot to measure against, so it scores zero for growth and climbs at its 15-minute re-scan. Fresh tokens are scored conservatively, which is the correct direction to be wrong in.

05 Verdict bands

BandRangeBehaviour
🎯 LOCK-ON80 – 100Ping fires immediately
📡 TRACKING60 – 79Watchlist; pings if it improves
🌫 DRIFTING40 – 59Visible in the terminal, silent
💀 GHOST0 – 39Flagged dangerous, warn-listed

Tokens are re-scored at 15 minutes, 1 hour, 6 hours, and 24 hours. A score at birth and a score after six hours of holder data are two different truths, and pretending otherwise is how people get rugged.

06 Replay & trust

Every score carries an inputsHash: a SHA-256 fingerprint of the exact inputs that produced it.

// the same inputs always produce the same number
const a = radarScore(inputs)
const b = radarScore(inputs)
a.score === b.score        // always true
a.inputsHash === b.inputsHash  // always true

When the archive ships, every LOCK-ON call is recorded with its outcome at +24h and +7d — wins and losses. The application role that writes those rows holds INSERT and SELECT, and holds neither UPDATE nor DELETE.

Deleting a bad call is not a policy we chose against. It is a capability the database grant does not give us. Losses that stay visible are the only thing that makes wins believable.

07 Outlaw Registry

Smart-money detection exists in no other product on this chain. HOODRADAR builds it from raw trader data. A wallet enters the registry on evidence:

  • at least 5 profitable exits
  • a win rate above 60%
  • at least $10,000 realized

A convergence signal fires when two or more registry wallets, with no shared funding ancestry, independently enter the same fresh token within minutes of each other. It is evidence that several proven traders agree, never an instruction.

Registry farming is a real risk. A well-funded adversary could manufacture wallets that clear the bar. Wallets trading predominantly against each other are clustered and excluded, and convergence requires no shared funding history — but the risk is not zero, and it is stated plainly rather than hidden.

08 API

Base URL: https://radar.hoodradar.fun/api/v1. Read endpoints are public with a delayed feed; real-time and the registry require an Overwatch-tier key.

MethodEndpointReturns
GET/score/:caFull scorecard, breakdown, inputs hash
GET/radarLive ping feed
GET/trendingScored trending list
GET/outlawsRegistry + convergence feed · Overwatch
GET/track-recordAppend-only call archive
POST/webhooksSubscribe to LOCK-ON / convergence · Overwatch
GET/healthStatus + per-endpoint data freshness
# score any token, no key required
curl "https://radar.hoodradar.fun/api/v1/score/0x020bfc…18b4"
# → { "symbol": "CASHCAT", "score": 71, "band": "TRACKING", … }
The first five projects on Robinhood Chain that integrate receive free Overwatch keys. No contract, no revenue share.

09 Tiers

Every tier returns the identical score. The tier changes when you see it, never what you see.

TierHoldUnlocks
🔍 ScoutfreeRadar with 10-min delay · 10 scans/day
⚡ Operator2,000,000Real-time pings · unlimited scans · custom alerts
🦅 Overwatch10,000,000Outlaw Registry · API key + webhooks
⚔️ Warlord50,000,000STRIKE auto-trade · firehose stream

Overage beyond an API quota is billed in $HOODR and burned. Tier balances sit locked in wallets, off the market, for as long as a user holds their tier.

10 FAQ

Is $HOODR live?

No. It has not launched. There is no presale, no whitelist, and no allocation. Anyone offering you one is a scammer. The contract will be announced only from @hoodradar_rh and t.me/hoodradar_rh.

Does connecting a wallet risk my funds?

No. The terminal reads your $HOODR balance through a read-only signature and discards it. It never requests a private key or a seed phrase, and it never constructs a transaction. HOODRADAR takes no custody at any phase.

Can the score be wrong?

Yes. A token can pass every check, score in the eighties, and still go to zero. The Radar Score is a heuristic, not a guarantee, and every call — including the losses — will live permanently in the public archive.

Why is holder growth zero on new tokens?

Because we have no prior snapshot to measure against yet. It is unmeasured, not bad, and it fills in on the token's next re-scan.

Radar Score is an automated heuristic rating, not financial advice. Trading small-cap tokens carries extreme risk. HOODRADAR is independent and unaffiliated with Robinhood Markets, Inc. DYOR.