Methodology

• StonkFun supplies the token universe: which tokens exist, launch details, graduation status, quote assets, transfer fees, burns, lifetime reward totals, revenue history, and the flywheel buyback roster. It is the only source of what a Stonk token is; it is no longer the source of market data.
• Birdeye supplies every market figure: prices, market cap, liquidity, 24h volume for the hundred largest tokens (the feed’s figure remains for smaller ones), hourly volume for the hot strip, candles, the largest holders and their trade history, the top traders by PnL, and each wallet’s realized and unrealized PnL in a token. Results are cached site-wide: token stats for five minutes, volumes for fifteen, holder reports for an hour.
• Solana via Helius supplies only the transaction stream of the StonkFun payout wallet that feeds the dividend index: a webhook for live payouts, an hourly reconciliation, and a gap check against the chain every ten minutes.

Anything OMEN cannot verify is shown as “—” rather than estimated.

StonkFun pays every dividend from one wallet, 5KXDF6QnqhBj72hDtJNkkpFaQVUfbFXNybMsp3DiK6tD, as batches of quote-asset transfers to holders. OMEN indexes that wallet itself instead of relying on StonkFun’s aggregates. Each payout transaction is stored with its quote asset, amount, USD value at the quote price when it was ingested, and recipient count, and is attributed to the Stonk token whose holders received it. Attribution uses StonkFun’s recent-distribution feed when it lists the signature; otherwise the token that every sampled recipient holds among the tokens paying in that quote asset.

• Dividends (24h) is the sum of indexed payouts over the last 24 hours. While the index is younger than a day the label shows the hours covered and the daily rate is scaled from that window. Until at least ten minutes are indexed the figure is shown as “—”; nothing is averaged or estimated in its place.
• Window: the 1h / 12h / 24h / 7d / All time selector on the ecosystem and tokens pages sets the span for dividends, APR, and flywheel buybacks. 1h, 12h, and 24h come from exact payout timestamps; 7d from hourly totals; All time from StonkFun’s lifetime figures. If the index is younger than the window, the label shows the hours actually covered.
• Per wallet: every payout parsed from chain is also stored per recipient, so the top-holders table shows what each wallet received from that token over the last 30 days and in total. Amounts are valued pro rata at the payout’s quote price; payouts not yet attributed to a token are counted for the wallet but not for the token until attribution catches up.
• APR is a trailing rate measured on a three-day window: dividends the index saw over the last 72 hours, scaled to a day, × 365, ÷ the value of the holdings that actually receive payouts. Tokens parked in pool and bonding-curve vaults, in the distributor and in frozen accounts earn nothing, so they leave the denominator (the largest accounts are inspected and any owned by a program-derived address or a platform wallet are excluded). Until that supply is known for a token the rate is on fully diluted market cap and says “on FDV”. A token younger than three days is annualised over its whole life and marked “young”. Every rate is also translated into dollars a day on a $100 holding, and no market-cap floor is applied: small tokens stay in the ranking.
• APY is what reinvesting every dividend into the token (a loop) earns at that APR held constant: (1 + APR × f ÷ 365)³⁶⁵ − 1, where f is the share of a dividend that becomes new eligible balance after the token’s transfer tax, 0.5% slippage and a 0.25% swap fee. It applies to a holding only when a strategy runs on it; the flag next to the figure says whether it does for the connected wallet. For tokens younger than three days the APY is shown as “more than” the daily compounding of a 100% APR, because the rate is not yet trustworthy enough to compound.
• APR (modeled) is shown on token pages for cross-reference. It follows the method used by dashboards such as StonkFun Payout Radar: 24h DEX volume × the token’s transfer-fee rate is taken as one day of rewards, then annualized on market cap. It is an estimate of what the tax should generate, not what holders received; the measured figure is what the payout wallet actually sent. The two diverge when volume is bursty, when part of the tax is retained as an operating fee, or when payouts lag collection.
• Next expected payout groups recent payout transactions into distribution events (transactions within three minutes of each other) and projects the median gap between the last events forward from the latest one. A schedule that has already slipped shows “any moment”.
• Dividends paid (all time) on the ecosystem page is StonkFun’s lifetime total; per-token “paid (all time)” is StonkFun’s figure as well, because the on-chain index starts from the day it was switched on.

• Top 10 tokens ranks by market cap. The flywheel columns show how much of StonkFun’s revenue share has bought back and burned each token: StonkFun directs a fixed share of platform revenue into buybacks of its top tokens by market cap and burns what it buys.
• Highest APR ranks every paying token on the three-day trailing rate with no size floor; “young” marks tokens whose rate covers less than three days. Most dividends paid ranks on the selected window.
• Hot stonks in the top bar are, among the twenty tokens with the most 24h volume, those with the most buy plus sell volume across venues in the last hour, from Birdeye, refreshed every ten minutes.
• Market cap and newest sorts in the directory query the full StonkFun directory. Volume, APR, dividend, price, and 24h change sorts rank the top 1,000 tokens by market cap locally, so the order always matches the numbers shown.
• Quote assets: pairing counts, paired market cap, and paired volume cover the top 1,000 Stonk tokens by market cap. “Top performing” is the quote asset’s own 24h price change.

Token pages list the 20 largest holders of a token, token accounts merged per owner wallet, with balance, share of supply, and USD value, alongside the top traders by realized and unrealized PnL that still hold the token and the wallets paid the most in dividends. PnL comes from Birdeye’s per-wallet ledger of DEX trades in the token: realized profit, the average cost of what is still held, and the number of trades. Unrealized PnL is recomputed at render time as (current price − average cost) × balance, so it follows the live price rather than a cached snapshot.

• Holder reports are cached for an hour and the merged table for thirty minutes per token; both load in the browser after the page renders, so crawlers never trigger them.
• StonkFun’s own operating-fee and swap-routing wallets are left out of every holder and recipient ranking.
• Wallets that never traded the token on a DEX have no cost basis and show “—” for PnL. The largest accounts often include pools and exchange wallets.

The Dividends page shows the connected wallet’s Stonk tokens only: those it holds (from Birdeye’s wallet token list) and those that paid it, even after they were sold. For each: what it paid this wallet over 24h, 7d, 30d, and all time from the payout index; the token’s APR and APY as defined above, or, when the yield index has no figure for the token, what the position itself earned in the last day annualised on its value; the estimated daily income that rate implies; the next expected payout from the token’s cadence; and the wallet’s PnL. Expanding a token lists the payouts it sent this wallet, five at a time. Two Birdeye calls per wallet (token list, then PnL for every Stonk token in one request), cached for five minutes; nothing is fetched unless the page is open with a wallet connected.

• Candles come from Birdeye at 15-minute (24H), 4-hour (7D), and daily (30D, 90D, 1Y) resolution, in USD, as market cap or price, cached for two, ten, and thirty minutes respectively. The axis switches to logarithmic automatically when a token moved more than about 6× inside the window. Charts are rendered with TradingView Lightweight Charts.
• Pages refresh their data every three minutes while visible and stay in the browser’s cache for three minutes, so switching between sections does not refetch. Every paid lookup is cached site-wide in Postgres and refreshed by one server at a time.