sources & method
Methodology
The glossary defines the terms used below: bonding curve, graduation, migration, creator reward, platform fee, LP fee, fees and revenue, settlement layer, lower bound, sample, survival, HHI.
Where the numbers come from
MemeFees sources platform fee and revenue figures from DefiLlama's fees dashboard. Market share and concentration are calculated from those figures. Source publication, collection and page refresh are separate steps: loading a page does not make an older observation current. Unavailable values are shown as a dash.
Fee and volume requests use a five-minute cache; the upstream reporting cadence varies by adapter. Listed coin quotes refresh every ten minutes, and the recent-launch feed refreshes every minute. By default, the token lists on New launches and Graduating hide tokens with measured liquidity under $1.0K and 24-hour volume under $1.0K or not reported. Each list states how many tokens it hides and can show them. Tokens without a liquidity measurement stay listed. Launch counts, statistics and the JSON feed include every token. Historical and active-trader series retain their reported UTC dates. Check the observation time and coverage note beside each metric, the endpoint catalog, and source status for delays and collection gaps.
Accounts, wallets and token creators
A total account count requires an explicitly reported cumulative count of registered platform accounts. The connected sources currently do not publish this for the profiles we cover. We show “Not reported”, not zero. Daily active wallets, token holders, token creators and transactions measure different things. A person can control several wallets, and the same wallet can appear on several days or protocols; we never add those counts into unique users.
Pons publishes a Dune-backed analytics snapshot with unique token creators, launches, trading volume, protocol revenue and creator earnings. We retain its scope separately from DefiLlama’s V2-only volume. Daily windows end at the stated UTC date, require every calendar day, and do not replace explicit lifetime totals. Provider snapshot generation, underlying observation date and our retrieval time are displayed separately. This snapshot is cached for five minutes and withheld after 72 hours without a source update.
DefiLlama daily wallet and transaction series use a 30-minute cache and retain their source methodology. Some optional endpoints may be unavailable; an HTTP failure never becomes a zero transaction count.
Fees vs revenue
Fees are the total amounts paid by users of a platform (trading fees on bonding curves, pool fees, launch fees). Revenue is the share of those fees the platform itself keeps, after creator payouts, liquidity-provider shares, buybacks and similar. A platform with high fees but low revenue is typically sharing heavily with creators or LPs.
Coverage and interpretation
- pump.fun is tracked separately from PumpSwap, its post-graduation DEX. Our pump.fun figures are the launchpad side only, which understates the total ecosystem.
- FOMO uses DefiLlama's FOMO Wallet adapter (category "Trading App"). Its reported app-trading fees extend beyond launchpad activity, but do not represent a combined total for every FOMO product.
- LetsBONK runs on Raydium LaunchLab infrastructure; DefiLlama splits the branded frontend from the underlying protocol.
- The main scoreboard's fee share uses DefiLlama's Launchpad category plus the additional platforms we track, such as FOMO. Overlapping settlement layers such as Meteora DBC are excluded from the denominator. Shares use the selected rolling period and available fee data, rather than volume or user counts.
- Volume series exist for the platforms supported by the volume sources. Volume coverage is narrower than fee coverage. A missing value can mean unavailable coverage or a temporary source failure; it does not mean zero activity.
- We rank the launchpads DefiLlama tracks fees for. Launch platforms that only report TVL (ICO- and locker-style products like DxSale, or crowdfunding protocols) cannot be ranked by activity and are out of scope.
Historical market share on the statistics page uses complete UTC days for the platforms in our stored daily series, excluding overlapping settlement layers. It can differ from the main scoreboard because both the time windows and platform universe differ. Missing observations reduce coverage; the result describes the available tracked data.
MemeFees calculations and research
MemeFees uses reported fee data to calculate market share, fee concentration and per-chain totals. Platform APIs provide token-level observations. Our separate collection process records launches and graduations for the supported feeds and contracts. Platform profiles and comparisons add documented fees and launch mechanics.
Platform-level token data
Top-token tables on platform pages come from each platform's own public API: Jupiter's token API (which labels Solana tokens by launchpad), Raydium LaunchLab's list API (LetsBONK bonding progress), Zora's Coins API, the clanker.world API and four.meme's ranking API. Each table names its source. These are the platforms' own reported numbers, not our estimates; when an API is unavailable the section simply disappears rather than showing stale or invented data.
Wallet-level performance
MemeFees does not publish wallet profit-and-loss or win-rate rankings. The trader-position samples on platform pages measure activity within a token sample; they are not a record of individual wallet performance. Graduation and survival statistics use separate observations, with their definitions and coverage shown on those pages.
Curated facts
Launch costs, fee structures, graduation mechanics and creator-reward descriptions are curated from each platform's official documentation and announcements, and kept deliberately qualitative where platforms change parameters frequently. Fee schedules in this market change often. Always verify against official documentation for the current terms. Corrections and material updates are recorded in the changelog.
Corrections
Material corrections are recorded in the changelog with the date and affected information. Platform pages include source links and the recorded date of their fact checks.
Independence & disclosure
Rankings follow the reported metric and selected period. MemeFeesis independent of the platforms listed. Referral links are labelled beside the link and may earn MemeFees a commission; they do not affect the ordering. Crypto assets involve substantial risk. The site provides information, not financial advice.
Metrics catalog (own data engine, since 2026-09-03)
Every five minutes the MemeFees engine reads its configured launchpad feeds and factory contracts and writes observed results into its own database. Days are UTC. Each metric below states its formula, the platforms it applies to, and its caveat. The first full day of engine data is 2026-09-04; earlier days show only what was counted exactly (factory events since 2026-08-16) or DefiLlama series.
A separate job checks due follow-up measurements every minute. Provider failures are retried within the measurement window; observations arriving more than ten minutes late are excluded. Factory counters resume from each contract's last committed block. Recoverable gaps are processed oldest first, and affected historical daily totals are recalculated.
DexScreener supplies missing prices and metrics from the deepest matching pool. Jupiter can add Solana prices and holder counts. Every filled field retains its provider and observation time. These sources have finite coverage; missing values stay unknown.
The engine keeps row-level token data for 14 days, or 12 days when the database is near its size limit. Row-level data is the rows for notable tokens, their curve samples, graduations and follow-up measurements. The period starts at the last observation of a token, so a token that sources still list keeps all its rows. Older row-level data is archived as daily compressed files: one CSV file for each table and UTC day, with a checksum. The engine deletes a row from the database only after the archive file is written and read back. Before the engine deletes rows, it stores their daily values and adds their counts to permanent totals. Daily aggregates, lifetime counts and all-history survival figures stay in the database. The survival figures keep every measured market-cap multiple, so the published medians do not change when rows are deleted.
| Metric | Formula | Applies to | Caveat |
|---|---|---|---|
| Launches per day | Tokens first seen by the engine, bucketed by creation time (UTC). EVM pads: factory events counted by block cursor. | All pads | A day is complete only with evidence for each counting method. Factory events: the counted block ranges of every configured contract cover the whole UTC day. A range hole or a new checkpoint makes the day incomplete; a collection pause alone does not, because checkpoints never skip blocks. Virtuals counter: samples at most 15 minutes apart, no decrease, and no change across midnight. Feed samples (pump.fun, four.meme, Clanker, Zora, Raydium, Jupiter, Bags) are always lower bounds, because no feed is verified to list every launch. A plus sign marks a lower bound. /api/v1/health lists complete days and reasons per platform. |
| Graduation rate (24 h) | Tokens created on day D that graduated within 24 hours ÷ launches on day D. | Curve pads only | n/a for pool-from-start pads (Clanker, Zora, Pools, LetsCash, NOXA) and for days with an incomplete launch count. |
| Median time to graduation | Median of graduated_at − created_at per day. | Curve pads | Graduations whose exact time was not reported by the source are excluded. |
| Survival after graduation | Share of graduated tokens whose market cap at +1 h, +24 h, +7 d is at or above the market cap at graduation; median multiple. | Curve pads | Both the graduation baseline and the horizon observation must arrive within 10 minutes of the reported event or due time. Estimated graduation times and late observations are excluded. Missing pools and unavailable prices stay unknown. |
| Market share | A pad's daily fees ÷ fees of all tracked pads that day. Settlement layers (Meteora DBC) are excluded from the denominator. | All pads | Calculated from dated DefiLlama observations; missing reports and overlapping settlement layers are excluded as documented on the dataset page. |
| Curve stall rate | Tokens that first reached 80 % of the curve on day D and did not graduate within 24 h ÷ tokens that first reached 80 % on day D. A graduation also proves the 80 % point, which covers curves that passed 80 % between two samples. | Curve pads with curve samples | Published daily on /stats/curve-stall-rate. Day D closes after D+1 ends. Curves come from sampled feeds. A graduation the engine did not observe counts as a stall, so pads with sparse graduation feeds (four.meme) show an upper bound. Estimated graduation times after the 24-hour mark leave the token out. Stored daily because row-level token data leaves the database after 12 to 14 days (it moves to daily archive files); the series starts on 2026-09-04. |
| Liquidity decay | Median of liquidity at +24 h ÷ liquidity at +1 h, and of liquidity at +7 d ÷ liquidity at +1 h, over tokens that graduated on day D. | Curve pads with follow-ups | Both observations must arrive within 10 minutes of their due time, and the +1 h liquidity must be above zero. The +24 h value closes after D+1 ends, the +7 d value after D+7 ends. Published on /stats/liquidity-decay. |
| Volume decay | Median of the rolling 24-hour volume at +7 d ÷ the rolling 24-hour volume at +24 h, over tokens that graduated on day D. | Curve pads with follow-ups | Both are complete 24-hour windows after graduation. A median of zero means more than half of the cohort had no volume a week later. |
| Answer pages | One page per launchpad and metric, for example /stats/pons/daily-revenue. The first sentence states the value, the UTC day and the denominator. | Six launchpads | A page is listed for search only with 14 complete days for its metric. Launch counts and graduation rates need proven launch counts, so feed-counted pads stay unlisted and say "at least". Revenue and market share use dated DefiLlama days. Survival needs 5 measured graduates per day and 30 in total. Pons and Flap.sh have no graduation source, so they have no graduation-rate or survival page. |
Live pages: /stats. Engine health and per-source status: /status.
Data license & open API
MemeFees-produced datasets are offered under CC BY 4.0 with attribution and a link to memefees.com. Data obtained from third parties remains subject to the source’s terms. The scoreboard is available as JSON at /api/stats; the daily series (fees, volume, launches, graduations, survival, market share) as JSON and CSV under /data. If you build on the API, credit MemeFees with a link.
Citing MemeFees
Link to the page that contains the figure or statement you use, and credit MemeFees with the upstream source shown there. Platform statistics, documented fee schedules and editorial opinions are different kinds of information; a trading fee percentage is not the same measure as a platform's collected fees in USD.
- For a number, retain the platform, metric, unit, reporting window, chain and scope. Distinguish a rolling 24-hour total from a completed UTC day.
- Include the observation date and when you accessed it. API processing times such as generatedAt and evaluatedAt are not new observations; preserve each source's dated values.
- Keep missing values, partial coverage and delayed-source notes with the result. A sample of discovered tokens is not the total number of tokens on a chain.
- For a fee schedule or review, use the recorded editorial date and linked official sources. That date does not establish that a changing price or fee is current.
Public HTML contains the main text and tables. The site retrieval guide links to smaller Markdown references for platform facts, reviews and referral guides, plus dated JSON and CSV data. These references use the same published records and identify the canonical human page. They are available to readers, researchers and automated systems without an account or API key.
Methodology and coverage notes updated .