How we know what we publish.

On-chain supply is read directly from each token contract across 26 tracked chains via RPC — Alchemy (archival) on EVM chains, Helius on Solana, XRPL native on the XRP Ledger, Horizon on Stellar, and Algorand indexer. Snapshots run every 4 hours, scheduled at 00:05, 04:05, 08:05, 12:05, 16:05 and 20:05 UTC; the scheduler is best-effort, so a run can land later than its slot — the "as of" timestamp shown on each surface is the authoritative collection time. Aggregators are used only as cross-checks — they are never the canonical supply source.

For coins where an issuer treasury wallet is publicly identifiable (currently EURC, EURS, EURQ, EURR and EURT), circulating_supply = total_supply − treasury_balance. The treasury balance is read from the same RPC snapshot at the same timestamp. For all other coins the token contract's totalSupply already reflects circulating (mint/burn model), so no subtraction is applied. If a treasury wallet cannot be sourced, circulating_supply equals total_supply and a caveat is shown.

Peg price is sourced from direct exchange feeds via CCXT (CEX-traded coins) or DexScreener DEX pools only. There is no price-aggregator fallback: a coin with no direct market source has peg marked "N/A" / "No public source" — never filled from an aggregator and never assumed to be €1.00.

The Herfindahl-Hirschman Index shown on the directory page is the sum of squared market-share percentages across all active non-sunset coins. HHI = Σ(share_i²). A value above 2500 indicates high concentration. Computed in-page from the same circulating supply figures used for the ranking — no separate data source.

The four "Best For" scores (Trading, Redemption, Treasury, General Access) are 0–100 heuristics built from verifiable, source-cited issuer attributes plus live peg and liquidity data. They reflect a use-case fit, not a safety or quality rating. The highest scorer in each column is highlighted. Scores are pre-computed daily and cached; the on-the-fly fallback uses issuer attributes only when the daily row is absent.

Each coin carries one product-lifecycle state — announced, active, wind_down or inactive — describing the ISSUER/product status, not market activity. A legitimate operating coin stays "active" even with tiny supply, one holder or no market; a coin becomes "wind_down" or "inactive" only on documented issuer evidence (a published wind-down, discontinuation or shutdown), never inferred from low activity. Poor liquidity or no market is recorded separately as market/activity metadata.

A failed, partial or unavailable read is represented distinctly from an observed zero — it is never stored or shown as 0. A missing row in a series is a real gap (interpolated or shown as a gap, not a zero), and a chain leg that could not be collected is marked unavailable so a day’s total is flagged incomplete rather than silently under-counted. Known unreconstructable historical windows are documented as coverage gaps.

Supply snapshots are not real-time — the scheduled cadence is ~4 hours for supply and ~6 hours for peg, and when a coin’s collection degrades its last snapshot can be older than that; every surface stamps the “as of” time and /data-coverage lists per-coin data ages. Holder counts are daily snapshots with no time series. Activity (mint/burn, transfer volume) coverage varies by coin — early periods may be sparse. A missing row in a series is a real gap, not a zero.

Peg price must come from an independent market source — a live exchange feed via CCXT (for CEX-traded coins) or DexScreener pool data (for DEX-only coins). Issuer-stated par values ("1:1 EUR") are not ingested as peg data. A coin with no public price source is marked "Not applicable" or "No public source", never €1.00.

Peg deviation = (market_price − 1.00) / 1.00, expressed as a percentage. The rolling average uses a 7-day window of daily closing prices. Deviations are computed from direct exchange feeds, not from aggregators. The display peg band is ±1.00% (±100 bps): a coin whose latest deviation falls outside it is flagged out-of-band with a marker pointing to the coin page.

A peg value is treated as a CURRENT market observation only while it is fresh (within 48 hours — peg is a daily series collected several times a day). An older value is not shown as the current price: rankings, coin pages and the API render "—" / null and expose the age instead, while the value stays available as history. One freshness policy is used across every surface.

For DEX-only coins the peg is computed from a pool’s price_native (the base token priced in its quote token, at full precision) — an EUR-quoted pool gives the EUR price directly, a USD-family pool is converted via the EURC-implied EUR/USD rate. DexScreener’s rounded price_usd (≈3 significant figures) is never used as a peg observation on its own, because rounding made unrelated coins share bit-identical fake prices. A pool with no usable precise price is dropped rather than published as a false quote.

Issuer redemption commitments ("holders may redeem at par") are classified as Issuer-disclosed commercial claims and are never treated as market peg data. The peg field reflects what the market actually trades at, not what the issuer promises.

CEX price is sourced from direct exchange APIs via CCXT — only allowlisted venues. DEX price is sourced from DexScreener pool snapshots. Peg uses these direct sources only; there is no aggregator fallback. A coin with no direct market source is marked N/A rather than filled from an aggregator.

Transfer volume is the gross value of peer-to-peer token movements per UTC day, per chain. Mint (from the zero address) and burn (to the zero address) are excluded, as are self-transfers, so the figure reflects value moving between holders — not issuance or redemption. It is a value figure in the coin’s units (≈ EUR), not a USD or fiat conversion.

Issuance and redemption are tracked separately as mint/burn, never folded into peer transfer volume. Mint/burn is read from on-chain zero-address events for every coin — with one exception: EUROP’s mint/burn figures are supplied by Schuman’s reporting infrastructure (its issuance is not fully reconstructable from public on-chain events). EUROP’s transfer-volume series additionally counts client-facing issuance/redemption to and from external holders while still excluding issuer-reserve and reserve-internal movements; it is presented under the neutral "Transfer volume" label.

A coin’s daily transfer volume is the sum of its observed per-chain legs for that day. Each leg is one chain, measured independently (Etherscan/Alchemy getLogs on EVM chains, Helius on Solana, Hubble/BigQuery on Stellar, native indexers elsewhere). A coin’s total is only "complete" when every chain that was live on that date produced an observed leg; a failed or missing leg marks the day partial rather than being summed as a zero.

For coins whose supply bridges across chains, bridge mint/burn legs are handled so cross-chain movement is not double-counted as peer transfer volume. Where a coin’s bridge legs are not yet fully attributable, the affected days are marked partial and disclosed on /data-coverage rather than presented as complete.

Historical reconciliation uses the chains that were live for a coin ON THAT DATE, not the coin’s chain configuration today: a chain contributes only between its launch and its documented collection-end. This keeps early single-chain history from being back-stamped onto later multi-chain days and vice-versa.

A failed or unreachable read is recorded as unavailable/failed, never as an observed zero — so a collector outage cannot masquerade as "no activity". Days with an unavailable leg read as partial; genuinely unreconstructable historical windows are documented coverage gaps. Coverage status is exposed on /data-coverage and in the API.

Weighted average: 40% liquidity depth score (from coin_liquidity_metrics_daily); 25% peg stability score (from coin_peg_metrics_daily, lower deviation = higher score); 20% CEX listing count (5 or more venues = 100); 15% 24h volume (€1M or more = 100). Measures how practically tradeable the coin is on public markets.

Weighted: 30% direct redemption availability (issuer-document fact, yes/no); 25% MiCA status (emi_licensed = 100, under_review = 50, other = 10); 20% onboarding requirement (none = 100, KYC = 70, KYB = 50, institutional = 20); 15% peg stability; 10% redemption fee (0 bps = 100, higher fees reduce score, unknown fee scored neutral). Measures how accessible redemption is for a typical user.

Weighted: 35% MiCA status (same scale as above); 25% attestation cadence (monthly or daily = 100, quarterly = 70, annual = 50, ad hoc = 30); 20% contract controls (freeze + blacklist + pause capabilities, each present adds to the score — the compliance tooling a regulated treasury issuer is expected to have); 10% reserve report publicly available; 10% peg stability. Measures structural transparency of reserves and issuer accountability.

Weighted: 25% access type (retail = 100, business = 50, institutional = 20, whitelist = 10); 25% onboarding requirement (same scale as Redemption); 20% tracked chain count (5 or more = 100, scale linearly); 15% decentralization factors (no freeze capability + no blacklist + no upgradeable contracts, each adds 33 points); 15% peg stability. Measures how permissionlessly the coin can be used across networks.

Supply snapshots are scheduled every 4 hours; peg snapshots every 6 hours. That is the collection cadence, not a staleness guarantee — if a coin’s collection degrades, its last good snapshot is served with its true "as of" timestamp, and /data-coverage lists per-coin data ages. No data on this site is real-time.

Tokens bridged across chains can be counted on both the origin and destination chain if both are instrumented, inflating supply figures. We subtract locked bridge balances only where a canonical bridge lock address is publicly identified and sourced. For chains where this is not yet done, a caveat is displayed on the coin page.

Not all chains a token is deployed on are fully instrumented. The coin page shows "N of M registry chains" where N is the count with active snapshot collection and M is the total deployment count. Supply figures for partially-covered coins are lower bounds, not totals.

Mint/burn and transfer volume history varies by coin — early periods may be sparse or absent. A missing row in the series is a real data gap, never substituted with a zero. The data-coverage page shows per-coin, per-dataset completeness percentages.

Holder counts are daily snapshots with no continuous time series. They are sourced only where a clean per-chain indexer exists; coins with no holder data show "Not available" rather than an estimate.

DefiLlama is used as a supply cross-check only, never as a canonical supply or peg source. Where our on-chain figure diverges from an aggregator's, we publish our sourced figure and note the discrepancy on the coin page if material.