Documentation

Reading the tape

TAPE is an on-chain forensic scanner for Solana. It reads a coin's complete transaction history and grades every wallet already in it — so you know who you're trading alongside before you ape.

On this page Overview · How it works · The score · Signals · Grades & tags · Tech stack · Safety · API · $TAPE · Limitations · FAQ

Overview

Most launchpad tokens are dominated by wallets that flip, dump, or rug — and by the time a chart looks good, those wallets already have their position. TAPE surfaces that read for you. Paste a token mint or a wallet address and TAPE pulls the on-chain tape, classifies behaviour, and returns a single 0–100 score plus behaviour tags, with a link to Solscan on every grade so you can verify it yourself.

Read-only. TAPE never connects a wallet, never asks you to sign, and never touches funds.
Transparent. The score is a sum of six published signals — no black box.
Solana-native. Built directly on mainnet transaction data via Helius.

How it works

Every scan runs the same pipeline:

Read the tape. For a coin, TAPE fetches current holders; for each wallet it pulls recent parsed transaction history — the full arc of buys and sells, not a balance snapshot.
Extract features. From that history TAPE derives behavioural features: distinct coins touched, buy/sell counts, fast round-trips (buy then sell within ~12 minutes), sell share, and wallet activity age.
Score & grade. Features feed the scoring engine, which returns a 0–100 score, a primary grade, and behaviour tags.
Open the receipts. Every graded wallet links to its Solscan page so the underlying tape is one click away.

Coin vs wallet scan

Coin scan takes a token mint. TAPE grades the top holders and shows the distribution — how many are clean, flippers, dumpers, or fresh — plus the share of scanned supply held by flippers and dumpers, and an aggregate tape grade for the coin.

Wallet scan takes a single wallet address and returns that wallet's score, tags, and signal breakdown across every coin it has touched.

Read it as context A heavy tape is not a verdict. A coin loaded with flippers can still run if the narrative is strong enough — TAPE tells you who's in the room, not what price will do.

The TAPE score

Every wallet starts at a base of 60. The six signals below add or subtract points, and the result is clamped to 1–99. Higher means cleaner, higher-conviction money; lower means extractive or unknown.

// scoring, in one line
score = clamp(60
  + hold_conviction   // up to +22
  + track_age_depth   // up to +16
  + spread_diversity  // up to +10
  - flip_rate         // up to -30
  - sell_pressure     // up to -16
  - fresh_wallet_risk // up to -20
, 1, 99)

Confidence is reported separately and scales with sample size — a wallet with only a handful of transactions is graded with low confidence, because there simply isn't much tape to read.

The six signals

+22Hold convictionRewards wallets that hold positions instead of flipping into the first candle. Driven by a low fast-flip rate across real round-trips.

+16Track age & depthEstablished wallets with real history score above brand-new accounts. Combines activity age and transaction depth.

+10Spread & diversityDiversified wallets look different from coordinated, single-coin clusters. More distinct coins touched raises the signal.

−30Flip rateThe heaviest penalty. Buying then dumping within minutes, repeated across coins, drags the score down hard.

−16Sell pressureWallets that are net sellers into their own positions read as extraction, not belief. Only bites above a balanced buy/sell ratio.

−20Fresh-wallet riskBrand-new, thin-history wallets are unknowns. Flagged until they build enough tape to earn a real read.

Grades & tags

The score maps to a primary grade:

Grade	Score	Meaning
CLEAN	70–99	Conviction money — holds, diversified, real history.
MIXED	45–69	Some flipping or selling, but not toxic. Read the tags.
FRESH	varies	Too new or too thin to grade with confidence.
TOXIC	1–44	Extractive pattern — heavy flipping and/or dumping.

Tags describe the specific behaviour behind the grade:

Tag	What it means
DIAMOND	Low flip rate, real round-trips, aged wallet — holds through volatility.
STEADY	Balanced behaviour, nothing extractive.
FLIPPER	High share of fast buy-then-sell round-trips.
FARMER	Touches many coins with a high flip rate — volume farming.
DUMPER	Net seller with a high sell share into positions.
FRESH	New wallet, thin history — an unknown.

Receipts

TAPE shows its work. Every graded wallet in a result links directly to its Solscan account page. Nothing is asserted that you can't check against the chain yourself — the receipts are the point, and they're what separates a real read from a vibe.

Stack & data

Chain: Solana mainnet.
Data: Helius — holders via the DAS getTokenAccounts method, wallet history via the Enhanced Transactions API.
Backend: a small Node/Express server that holds the Helius key server-side, runs the scoring engine, caches results in memory (~2 minutes), and retries with backoff on rate-limit responses.
Frontend: a static page served from the same origin — the key never reaches the browser.
Explorer links: Solscan.

On data freshness Scans reflect near-real-time chain state. Popular results are cached briefly to stay within rate limits, so a coin scanned seconds ago may serve a cached read for a short window.

Safety & privacy

No wallet connection. TAPE never prompts you to connect or sign. It reads public chain data only.
No custody. It cannot move, hold, or access funds.
Public data only. Everything TAPE shows is already on-chain and visible on any Solana explorer — TAPE just reads and grades it.

API reference

The same engine that powers the site is available as a small JSON API. Point requests at your deployment's base URL.

Set at launch Replace https://your-app.replit.app below with your deployed backend URL.

Scan a coin

# POST /api/scan/coin
curl -X POST https://your-app.replit.app/api/scan/coin \
  -H "Content-Type: application/json" \
  -d '{"mint":"<TOKEN_MINT>"}'

// response (trimmed)
{
  "subject": "<mint>",
  "holderCount": 412,
  "rows": [
    {
      "holder": { "owner": "<wallet>", "pct": 8.4 },
      "grade": {
        "score": 81,
        "grade": "CLEAN",
        "gclass": "clean",
        "tags": [ { "t": "DIAMOND", "c": "clean" } ],
        "confidence": 88,
        "comp": [ /* six signal components */ ],
        "f": /* raw features */
      }
    }
  ]
}

Scan a wallet

# POST /api/scan/wallet
curl -X POST https://your-app.replit.app/api/scan/wallet \
  -H "Content-Type: application/json" \
  -d '{"address":"<WALLET>"}'

// response
{ "subject": "<wallet>", "grade": { /* same grade object as above */ } }

Health check

# GET /health
curl https://your-app.replit.app/health
// -> { "ok": true, "service": "tape", "cached": 3 }

Errors return an HTTP status with a JSON body: { "error": "..." }. Common cases: 400 missing field, 404 no holders / no transactions, 429 upstream rate limit (retry after a moment), 500 upstream failure.

$TAPE & burn-to-access

$TAPE is the access key to the scanner. Instead of subscriptions or logins, you burn a small amount of $TAPE to unlock scanning for 24 hours. Every burn permanently removes supply, so usage tightens the float over time — the flywheel.

Action	Cost
Wallet scan · 24h	1,000 $TAPE
Full coin scan · 24h	10,000 $TAPE

Set at launch Contract address: CA — coming at launch. Burn amounts above are placeholders — tune them to your supply before launch.

The token's role is utility — access to a tool — not a claim on revenue or profit. That framing is intentional and should carry through all launch copy.

Limitations

Heuristic v1. The scoring signals are behavioural heuristics with sensible defaults. Thresholds are tuned against live data over time.
Bundler & rug detection are approximations. True same-slot bundle clustering and liquidity-pull rug detection require additional enrichment (token creation time, LP events) and are on the roadmap — TAPE flags behaviour it can measure and does not fabricate signals it can't.
Sampling. Coin scans grade the top holders and read recent history per wallet; extremely deep histories are truncated, which is reflected in the confidence score.
Not a price predictor. TAPE grades who holds a coin, not where it's going.

FAQ

Do I need to connect a wallet?

No. TAPE is read-only and never prompts for a wallet or a signature.

What chains does it support?

Solana. Support for other venues may follow.

How current is the data?

Near-real-time from mainnet, with brief caching on popular scans to respect rate limits.

Why did a scan fail?

Usually an invalid mint/wallet address, a token with no holders, or a momentary upstream rate limit. Check the address and try again.

Can I build on the API?

Yes — see the API reference. Be mindful of your own rate limits and cache aggressively.

Is a high score a guarantee?

No. Scores describe past behaviour of current holders. They're a strong signal, not a promise — always do your own research.

Disclaimer

TAPE provides on-chain behavioural analysis for informational purposes only. It is not financial, investment, or trading advice, and nothing here is a recommendation to buy or sell any asset. Behavioural signals are probabilistic and can be wrong. Trading crypto assets — memecoins especially — carries a high risk of total loss. Always do your own research and never risk more than you can afford to lose.