How ORSOL reads a token

ORSOL is a scanner for Solana tokens. Paste a token's mint address and it tells you how risky the contract is, who really holds the supply, and how the market is behaving, all on one page within a few seconds.

Most scanners stop at “the top 10 wallets hold 40%”. That number can't tell you whether those ten wallets are ten different people or one person with ten wallets. ORSOL goes further: it replays the token's launch and follows the money that funded the biggest holders, so it can point out bundles, snipers, dev exits and wallet clusters.

1. 1

   Copy the mint address

   The mint (also called CA or contract address) is the token's ID on Solana, 32–44 characters long. You'll find it on Dexscreener, Jupiter, pump.fun or the project's own site.
2. 2

   Paste it and press Scan

   Use the search box on the home page or at the top of any results page. Wallet addresses and pool addresses won't scan, only token mints.
3. 3

   Read the results

   The score and the contract checks appear first. Wallet forensics load a few seconds later and can update the score; the AI read appears alongside.
4. 4

   Share it

   Press Share on X under the score. A preview shows the exact post first: a short line written from your scan (the score and the findings that matter), the CA, links to the token on Dexscreener and GMGN, and the scan link /t/<mint>, which unfolds into a card with the score. Post it on X in one tap, or copy it for Telegram and Discord.

A results page is laid out from “quick answer” to “full detail”:

Each signal carries a colour so you can skim:

Every scan runs the same four stages. The first one answers in a few seconds; the second runs in the background.

1. Read

We read the mint account from Solana to get its authorities and Token-2022 extensions, resolve the 20 largest holders and work out which of them are pools or programs, and decode the pump.fun bonding curve if the token has one. In parallel we fetch market data, Jupiter's token info and the LP lock status, then the hourly candles for the main pool.

2. Trace

We pull the token's first few dozen transactions and record who bought, in which slot, and how much. Then, for the dev, the largest holders and the launch buyers, we look up where each wallet's first SOL came from. Details in Wallet forensics.

3. Score

The checks are grouped into four pillars, each scored 0–100, weighted into one total, then limited by any hard cap that applies.

4. Explain

A four-sentence summary is written from the finished numbers. It can only quote numbers that are already on the page.

The score is a weighted average of four pillars, each scored 0–100. Contract, Wallets and Fundamentals start at 100 and lose points for every risk they find. Technicals starts at 50 and moves up or down with the trend. Either way, every point lost shows up as a signal on the results page.

Technicals needs at least 30 hourly candles. With fewer, that pillar is left out and the other three are re-weighted to make up 100%.

Verdicts

Some problems are serious enough that no chart or volume should hide them. When one of these applies, the final score can't go above the cap, no matter how well the other pillars did. The lowest applicable cap wins.

This is the part that sets ORSOL apart. It answers a simple question: of the wallets holding this token, how many are really connected to the launch or to each other?

Insider supply

The headline number adds up what the dev, bundle wallets, snipers and cluster wallets still hold today. Wallets that bought early and already sold don't count. Fresh wallets are shown separately and are not counted as insiders.

Exchanges are excluded

Thousands of unrelated people withdraw from Binance, Coinbase, OKX, Bybit or Kraken. Wallets funded from known exchange hot wallets are never grouped into a cluster.

Confidence

Authorities and metadata are read straight from Solana. LP status comes from RugCheck and the sell route from Jupiter.

Token-2022 extensions

Newer Solana tokens can carry extensions. Most are harmless; these are the ones we flag:

Tokens launched on pump.fun start on a bonding curve instead of a normal pool. We read the curve account directly from Solana and show how far it is to graduation.

The live list on the home page tracks the 40 newest pump.fun launches over a Solana WebSocket subscription, so progress moves the moment a trade lands on-chain. It shows the 12 most advanced active curves; drained curves (tokens sold but almost no SOL left) are left out.

Below the score is a four-sentence summary: what the market is doing, the biggest risk with its number, one thing to watch next, and a bottom line that matches the verdict.

It is written by Claude from the finished snapshot, not from raw data, and it is told to only use numbers that are already on the page and never to give buy or sell advice. If the model is unavailable, a rule-based summary is shown instead, so this panel is never empty. The label above it says which one you're reading.

Anything that decides whether a token is safe (authorities, holders, the curve, the launch) is read straight from Solana. Third-party APIs supply market data, and we only lean on them where the chain can't answer.

Live status

A real request to each provider, made by our server right now. Nothing on this panel is hard-coded.

What comes from where

On the home page

“Each scan” means fetched when you scan, then reused for 45 seconds so a token everyone is scanning doesn't hammer the upstream APIs. If a source fails, the scan carries on without it and the Sources panel on the result marks it unavailable.

The same data that powers the site is available as JSON. No API key is needed; each endpoint is rate-limited per IP address, and responses are cached for the time shown, so calling again sooner returns the same data.

All endpoints are GET.

curl https://orsol.app/api/v1/token/JUPyiwrYJFskUPiHa7hkeR8VUtAeFoSYbKedZNsDvCN

{
  "mint": "JUPyiwrYJFskUPiHa7hkeR8VUtAeFoSYbKedZNsDvCN",
  "scoreVersion": "1.0.0",
  "token": { "name": "Jupiter", "symbol": "JUP", "program": "spl-token" },
  "market": { "dex": "orca", "liquidityUsd": 619012, "marketCap": 809230000 },
  "pillars": [
    { "key": "risk", "weight": 30, "score": 90, "signals": [
      { "label": "Mint authority", "value": "Revoked", "tone": "good", "source": "Solana RPC" }
    ] },
    { "key": "distribution", "weight": 30, "score": 25 },
    { "key": "fundamental", "weight": 20, "score": 75 },
    { "key": "technical", "weight": 20, "score": 15 }
  ],
  "overall": { "score": 53, "verdict": "Neutral", "caps": [] },
  "sources": [{ "name": "Solana RPC · mint", "ok": true, "ms": 114 }]
}

const live = new EventSource("https://orsol.app/api/v1/live");
live.addEventListener("curves", (e) => console.log(JSON.parse(e.data).rows));
live.addEventListener("market", (e) => console.log(JSON.parse(e.data).rows));
live.addEventListener("sol", (e) => console.log(JSON.parse(e.data).usd));

Errors return JSON with an error message: 400 for an invalid address (or, on candles, a bad pool or tf), 404 for a valid address that isn't a token mint, such as a wallet, 429 when rate-limited (wait the seconds in the retry-after header) and 502 when an upstream source is down.

A scanner that rates its own token has a conflict of interest. ORSOL never gives its own token a score or verdict. Its page shows the raw checks with a clear “team token” label instead.

• Wallet forensics are heuristics and can be wrong in both directions. Treat them as leads to check, not proof.
• Tokens with a very long history can't always be traced back to launch. Bundle and sniper checks are then skipped and confidence drops.
• Home page prices update every 3 seconds and curves on every trade. A token's results page is a snapshot that can be up to 45 seconds old.
• A clean scan doesn't make a token safe. New risks can appear after you scan.