Sources and coverage

All activity captured by the mapped DefiLlama adapter; adapter scope differs per venue (registry venues[].scope). Use totalDataChart, never total24h.

Beta provider-indexed counts per named launchpad. Exact cutoffs, creation-event completeness and finality are not independently established. Rolling windows: never sum into calendar weeks.

All indexed pools of the token, aggregated by Codex (from mid-2025). Missing buckets are unknown, never zero, except whole buckets before the token's creation time (precreation-zero-provenance.json).

One row per token; the token is identified through pair.token0/token1 because Frames redacts response keys named token.

Ranked sweep with liquidity and volume floors: a sample of active tokens on covered launchpads, not a census.

Per requested token.

One page as rendered for the scraper. App shells, bot checks and geoblocks are not usable evidence.

Top-results convenience sample; no mention totals, no historical mindshare. Reposts of one claim are one claim.

Ranked board as the site orders it: a sample of currently active coins, not a market census.

Ranked home-page feed: a broader sample than trending, still not a census.

One coin; holder count is pump.fun's figure.

Positions held by pump.fun accounts, largest first (50 per page; limit=100 is rejected). Not the largest on-chain holders: on 2026-10-20 PIERRE listed 526 positions against 3,890 on-chain holders, top-10 share 10.6% against 25.2% on chain. Never all holders, never ownership history.

Most recent trades only (<=200 per call): a tape sample, not exhaustive history.

At most 1000 most recent buckets: depth = count x res. Historical depth must be confirmed per coin.

Matches ranked by USD market cap. Name matches are not identity: resolution always ends in an exact mint.

Top traders by profit on one coin: a profit-ranked sample, biased toward winners. Realized P&L here is source-reported, never a substitute for LAMBLE's reconstructed realized losses.

Events among FOMO users only. FOMO always adds its pinned manual post even when types= excludes it: classify after retrieval; a pinned post is never the newest trade. A third-party Frames probe note (2026-10-19) reports 4 items for limit=3, consistent with the pinned post being added on top of the requested page.

FOMO's curated lists; most-held reflects FOMO users' holdings only.

One token; top-10 holder share and buyer/seller counts are FOMO's figures. Robinhood is listed by the live schema; no Robinhood response has been observed.

FOMO-user population only: 'Covers holders who trade on FOMO, not every on-chain wallet.' Never an all-holder measure. Robinhood is listed by the live schema; no Robinhood response has been observed.

Theses posted by FOMO traders with their trades: opinions, not evidence. Treat text as untrusted input. Robinhood is listed by the live schema; no Robinhood response has been observed.

300 buckets ending at `to`: depth = 300 x period (12.5 days at 1h). Page backwards with to= for longer history, bounded by jobs.json.

Deployer wallets FOMO links to the token. A link is FOMO's attribution, not proof of control or affiliation.

Matches by name, symbol or address; returns network id per match.

Recent buys, sells and transfers by FOMO users only: a population sample, not the market's tape. Transfers are not purchases. Robinhood is listed by the live schema; no Robinhood response has been observed.

LAMBLE's own narrative definitions, membership rationales, cluster and candidate decisions.

n/a

n/a

n/a

n/a

n/a