Where it comes from

• Card facts: the card DNA handoff (version 603b6c46), every card read fact by fact: roles, what it does, what it answers and how, what it costs. The site reads it and never edits it.
• Scryfall: legality in every format, colors, sets, printings, images, mana symbols and set icons. Legality and new cards read October 22, 2026; printings October 22, 2026.
• Popularity: EDHREC's rank, from Scryfall's Oracle file of the build (the card facts' own copy only when Scryfall has none), shown on a card page as its rank in Commander decks, with that file's date. It counts Commander decks, so it leans toward cards Commander players use.
• Card names, text and art belong to Wizards of the Coast and are shown under its Fan Content Policy.
• Combos: every combo on Commander Spellbook, the community combo database behind EDHREC's combo pages, from its public daily export (combos.py). Each combo's cards, where they start, prerequisites, steps and results are Commander Spellbook's words; mtgstuff matches them to its cards, sorts them, and links each one back to its page there.
• In constructed: the MTGO decklists Wizards and Daybreak publish, through the open archive MTGODecklistCache, with archetype names from MTGOFormatData: Meta Lens's own files, pulled October 22, 2026.

Legality

Each format's status (legal, restricted, banned or not legal) is Scryfall's, from the legalities field of its Oracle bulk file read October 22, 2026; a restricted card counts as legal, one copy. Each format's name links to its page on the site. Under a banned or restricted row is the date it was banned or restricted and the source that says so, from the site's formats file (kept by hand, every entry with its source): a date the source gives only as a month shows as a month, a date that is only the first record of a list reads "on the list by", and a secondary or approximate one says so. In Duel Commander, Competitive Brawl and Tiny Leaders the row reads "Banned as commander", never "Restricted": Scryfall stores those lists as restricted. A ban in MTG Arena's Best-of-One only, and a ban as a companion only, show under a legal row with their dates. A card not out yet reads "Joins Pauper on release" only when it is printed at common in an expansion, core, Masters, draft innovation or Commander set coming out (Star Trek's commons, November 13, 2026). A fold under the list holds the formats added since October 18, 2026, and says how many of them the card is legal in: Pauper Commander as two rows, the 99 (Scryfall's paupercommander legality) and as commander (Scryfall's is:paupercommander list, dated), Gladiator, Penny Dreadful, Old School 93/94, PreDH, Alchemy, Competitive Brawl, Tiny Leaders, and the three formats read from another format's legality by their own rule: Canadian Highlander and 7 Point Highlander (Vintage legal or restricted) and Big Apple Highlander (Premodern legal or banned), less their own bans. Points show on those rows (Canadian Highlander's and 7 Point Highlander's, and X-Point Old School's on the Old School row), each with its list's date and a link to the list; 7 Point Highlander's row also says that a deck with no Reserved List card gets 8 points, not 7. A card on the Reserved List says so, from Scryfall's reserved flag, and a card on Magic Online only through its Magic Online printings names those sets. The formats file is read after the card has drawn, into lines already kept for it, so nothing moves. Brawl is the 100-card Brawl, called Historic Brawl until December 2023. Scryfall's "future" key isn't a row: it isn't Wizards' upcoming Standard (it reads Star Trek's cards and The Big Score not legal, though Star Trek joins Standard on November 13, 2026 and The Big Score stays until the next rotation); f:future still searches it, named Scryfall's Future Standard, with that note. Every row Wizards has dated legality for that is still ahead says when the card joins (Star Trek's cards: Standard, Pioneer, Modern, Legacy, Vintage, Commander, Oathbreaker and Duel Commander on November 13, 2026), once per row; Canadian Highlander and 7 Point Highlander follow Vintage's date and Big Apple Highlander Premodern's. Where the formats file's own list and Scryfall differ on the card, the row says both (Alchemy: Wizards' page still names Leyline of Resonance banned; Scryfall reads it legal), and Penny Dreadful says when Scryfall reads a card legal that the season's list, made before its set, doesn't hold (Reality Fracture's cards). Each status is shown as Scryfall gives it. One exception to Scryfall's reading: where Wizards has announced a card's legality and Scryfall's file doesn't carry it yet, the site follows Wizards from the day Wizards names, only while Scryfall still calls the card not legal there, and until that day says when the card joins. Today that is Mystery Booster Commander Edition's 60 new cards (collector numbers 1 to 60, Seluma and the other Rulebreakers among them): they join Commander on November 13, 2026, and Legacy, Vintage and Duel Commander on January 1, 2027 (Wizards’ page, checked October 21, 2026). Old School 93/94 is another exception: Scryfall decides it printing by printing (an Alpha Black Lotus is restricted there, its Vintage Masters reprint isn't legal), and its card-by-card file carries one printing's answer, so the site reads every paper printing's own answer from Scryfall's printing file and takes the card's status from them: restricted when any printing is restricted, else legal when any is legal, else banned when any is banned. A build whose printing file lacks those answers shows Old School as "By printing" and leaves it out of searches with a sentence. Arena's Best-of-One bans come from Wizards' banned and restricted page, since Scryfall's legality doesn't carry them; Captain's line comes from its founders' lists.

Its wording, then and now

The rules text is Scryfall's current Oracle wording. Under it, a fold shows the text as printed on each English printing where it differs from today's wording, with the sets it was printed in, and the card's name in ten languages, both from MTGJSON (MIT License, Copyright Zach Halpern), read only when you open it. Each build compares today's Oracle wording with the last build's, and a card whose wording changed says so, dated, with the wording before. The log began on the first build that had a copy to compare with; changes before that day aren't listed.

The card first. A card page shows the card's name, cost, type, rules text, legality, price and picture from a small file picked by its address, while the full card list (about 2 MB) loads behind it, then the rest of the page. If the network connects but doesn't answer, the page says so after 8 seconds.

Why it is or isn't Pauper legal

The verdict is Scryfall's: its Pauper legality, and its list of the cards that can lead a Pauper Commander deck. The fold under the legality list shows what it rests on, from the card's own printings. Wizards' rule: a card is legal in Pauper when any printing of it was released at common in a paper product or on Magic Online, and then every printing of it may be played. So the fold lists every common printing on paper and every common printing made only for Magic Online (Vintage Masters, the Masters Editions, Magic Online's theme decks and promos), each with its set and its day, and marks the ones that don't count and why: Renaissance and Rinascimento, a localized release; promos; tokens; printings that aren't tournament legal (Collectors' Edition, gold borders, oversized and playtest cards); an Unfinity card with the acorn stamp; a set not out yet. A common printing on MTG Arena only doesn't count either, and the fold says so when the card has one (Arena's own lowest rarity, from the Wildcard Planner's file). A banned card shows its ban date and the source. Checked against Scryfall's verdict on every card on October 21, 2026: none disagrees. For Pauper Commander's commander the fold lists the uncommon printings, on paper and on Magic Online only, quotes PDH Home Base's rule for the commander and for which printings count, and gives Scryfall's answer and the day the list was read; it doesn't mark printings itself, since Scryfall's list (kept with PDH's own team) counts some its written rule leaves out, such as The Zeta Set's uncommons.

In constructed

How often the MTGO lists Meta Lens carries play the card, per format: Standard, Pioneer, Modern, Legacy, Vintage, Pauper and Duel Commander, each shown only when the card is legal, restricted or banned there (a card banned lately can still sit in older lists). A list plays the card when it holds at least one copy, in the main deck or the sideboard. Every event kind counts (Challenges, Leagues, qualifiers and the rest), as on Meta Lens with nothing switched off. MTGO publishes only the top finishers of each Challenge and only 5-0 League lists, so these are shares of published winning lists, not of everyone who played. They count every published list in the window, not a sample, so no range is shown.

Windows are whole days ending on the newest list's day in that format: the last 30 days is that day and the 29 before it, the last 90 days that day and the 89 before, and the 30 days before are days 30 to 59 back.

share = lists playing it / lists in the window

change, in points = (share, last 30 days) minus (share, the 30 days before), times 100

average copies, main deck = copies in main decks / lists with it in the main deck

average copies, sideboard = copies in sideboards / lists with any in the sideboard

Sideboard only counts lists with copies in the sideboard and none in the main deck. In Duel Commander, MTGO files the commander in the sideboard, so the rows read "In the 99" and "As the commander", with no copy averages. The small line is the share week by week for the 8 weeks of 7 days ending on the newest list's day, oldest first, with each week's figure in its label. Archetypes playing it are the five with the most lists playing it in the last 30 days (the last 90 when none in 30), ties by their own list count, each beside the share of that archetype's lists that play it; archetype names are Meta Lens's. Formats are listed from the highest share in the last 30 days; formats where no list in 90 days plays it share one line, with their list counts. The files load only when the section is open and near the screen (on a phone, when it is opened), only for the formats the card needs.

When the source pauses. The archive's README says its lists were updated automatically each day and that the project is no longer being actively maintained; it last changed October 12, 2026, so lists end October 11. The section says the newest list's date and how many days before today it was, so a pause shows on the page.

What answers it

Cards whose facts say they deal with this card's types (creature, artifact, and so on), legal in the format you pick, grouped by how they do it: exile, destroy, bounce, and the rest, with counterspells as their own list. Answers this card shrugs off are left out: ones its indestructible, hexproof, shroud or protection stops (only protection the card has itself; an Equipment's protection belongs to the equipped creature), damage too small for its toughness, and answers whose restrictions it doesn't meet. A known ward cost is a cost to pay, not a stop. Listed by EDHREC rank, which counts Commander decks only.

Combos

Every known combo the card is part of, most played first. The count of decks is how many EDHREC decklists run all the pieces, as Commander Spellbook reports it. The bracket is Commander Spellbook's own tag for the lowest Commander bracket the combo fits (Exhibition 1, Core and Oddball 2, Powerful and Spicy 3, Ruthless 4), an estimate separate from the bracket reading in Decks. The kinds (wins the game, infinite mana, tokens and so on) are read from the result names, so one combo can be several kinds. Mana is the total mana value the combo needs each time around. Legality per format is Commander Spellbook's. Salt, where Commander Spellbook has votes for a combo, is its score of how much players say they dislike facing it, 0 to 4, shown with the number of votes; Commander Spellbook states no license for its data.

Same job, other cards

Weighted likeness of the two cards' facts, cut for mana value distance:

likeness = shared / (A + B - shared) x (1 - 0.06 x min(mana value gap, 5))

Shown when above 0.15, and by default only inside this card's color identity.

Prices

TCGplayer market prices in US dollars from Scryfall's printing file, read October 22, 2026 and dated at the bottom of the page: the range across every printing and finish, and each printing's own price. A market price is TCGplayer's figure from recent sales of that printing and finish, not an offer from any store. The same file prices decks and collections, so the numbers agree across the site.

On Magic Online only folds the printings made only for Magic Online (Vintage Masters, Tempest Remastered, the Masters Editions, MTGO promos), cheapest first, with each one's Magic Online price in tix from Scryfall, dated. No store sells them on paper; they let a Legacy, Vintage or Pauper deck be priced on Magic Online where no paper printing is there (Underground Sea, Mishra's Workshop). The paper printings' own tix are in Super Search (tix<1).

Prices and where to buy lists every paper printing with its nonfoil, foil and etched price (a dot means that finish was never made), how many copies of each printing your collection holds, and a link to buy it (an etched foil links to its own TCGplayer product). "Lowest price on file" is the lowest of those prices by the site's price rule: printings you can play in a paper game, serialized ones aside. The table marks each printing the rule leaves out and why, and names the language of a printing that isn't in English.

eBay sold listings

A plain link out, nothing more: the site reads nothing back from eBay. Under the prices, the link searches eBay's sold listings for the card's name as stores list it (a two-faced card by its front face, a split card or Room by its whole name, Fire // Ice) and the word mtg; picking a printing adds a second link that also names its set. The link uses eBay's sold and completed filters (LH_Sold=1 and LH_Complete=1 on ebay.com/sch/i.html), with no category filter, since the site couldn't confirm eBay's category number for Magic singles. A sold listing is what one buyer paid one seller, in that listing's condition. eBay may ask you to sign in to see sold listings: since July 22, 2026 it sends signed-out visitors to its sign-in page for these filters, as OpenWeb Ninja's guide to eBay's sold and completed listings describes (read October 18, 2026). mtgstuff isn't affiliated with eBay and the link carries no partner code.

Where the store links go

Links only: the site never asks a store for anything. Each link opens the store's own public page in a new tab, and none carries an affiliate or partner code. mtgstuff isn't affiliated with TCGplayer or Card Kingdom and earns nothing from them.

• TCGplayer, one printing: its product page, tcgplayer.com/product/<product id>, with the product id Scryfall publishes for that printing (the same page Scryfall's own links open); the foil link adds Printing=Foil. TCGplayer sells an etched foil as its own product, so an etched link goes to that product's id (Scryfall's tcgplayer_etched_id). A printing with no product id on file gets TCGplayer's search for the card instead.
• TCGplayer, every printing: its product search for the card's name.
• TCGplayer Mass Entry: tcgplayer.com/massentry?productline=Magic&c= with one entry per card, joined by ||, in TCGplayer's own share-link format: 2-<product id> for the exact printing priced on the page, or 2 Card Name to let TCGplayer match any printing. An etched foil goes by its etched product's id. A card with no price on a printing you can play goes by name. A very long list is split into parts.
• Card Kingdom: its catalog search for the card's name (every printing it stocks; Card Kingdom's printings can't be matched to Scryfall's one by one from here), and its deck builder, cardkingdom.com/builder, sent the list as a form with one 2 Card Name line per card.
• Names: stores list a split card (Fire // Ice, aftermath cards, Rooms) by its whole name and a two-faced, adventure, flip or prepare card by its front face, so that's how the site writes them. The copy-the-list button uses the same names for any other store.

Table mode

Only what the Magic Tournament Rules let you show across a table (section 2.11): each face's name, cost, type line, rules text and power and toughness, loyalty or defense, from Scryfall's Oracle data, and the rulings Scryfall marks as Wizards' own. No prices, popularity, combos, links or tags. Turned around, the same text is rotated 180 degrees for the player across from you.

Every store's price

Beside the TCGplayer market price the site uses everywhere (from Scryfall), the card page and Collection show store prices from MTGJSON's AllPrices (MIT license), dated: TCGplayer retail as MTGJSON reads it, Card Kingdom's asking price and its cash buylist, Mana Pool's asking price and Cardmarket's price in euros, for the normal and the foil finish (etched has none). MTGJSON gathers these from the stores' public feeds, and the stores' own terms for that data aren't published; the site never reads a store's own pages. Card Kingdom pays is its cash buy price divided by the market price of the same printing and finish. The flag shows when the highest of TCGplayer retail, Card Kingdom and Mana Pool is more than 1.5 times the lowest. The stores' prices are for near mint copies and are shown as they are; only Collection's value by condition, off unless you switch it on, multiplies them by a condition share.

In Limited

From 17Lands' public game data for Premier Draft (CC BY 4.0), summarized by the site into counts of games and wins and nothing else. A card counts in a game when it was in the deck, in the opening hand, drawn later, in hand at all (the opening hand or drawn: GIH) or in the deck but never seen. Every rate is wins over games, with its 95% Wilson range; improvement when drawn is the win rate in hand minus the win rate when never seen, in points, with Newcombe's range. Color pairs need 200 or more games in hand. The 60% group is 17Lands' own bucket of each player's overall game win rate. 17Lands users win more than the average player, so compare within a set, not with 50%. The average pick is from 17Lands' draft data.