Making KeyForge proxies is not the same as building a deck in most trading card games. You are recreating one known Archon deck, including its exact card versions and printed bonus icons, rather than selecting 36 individually desirable cards.
A KeyForge deck contains 36 main-deck cards: 12 cards from each of three houses. It also has an Archon identity card with the deck list on its reverse. That identity card and any applicable reference components stay outside the shuffled deck. The official KeyForge Master Rulebook defines this deck structure and the game's current component rules.
TLDR
For a casual recreation, copy all 36 cards from one known deck exactly. Preserve duplicates, houses, card versions, enhancements and unusual inclusions. Make a separate Archon identity card, add any set-specific reference cards the deck needs, then use opaque sleeves with consistent backing cards so every card handles the same way. Do not bring the recreated deck to an official tournament: Ghost Galaxy's tournament rules prohibit proxy decks and cards.
Recreate a deck; do not construct one card by card
KeyForge's central idea changes the entire proxy workflow. An ordinary Archon deck arrives as a complete generated object. The useful starting point is therefore a specific deck list from an Archon identity card or another record you can verify, not a pile of favorite cards from three houses.
Even KeyForge formats involving deck construction do not become unrestricted card-by-card building. In Sealed and Alliance contexts, construction is based on complete 12-card house pods from eligible decks, with three different houses represented. That is a separate task from recreating a normal Archon deck for testing.
This distinction matters because a casually invented list will not tell you how the known deck actually performs. Its weaknesses, awkward draws and odd combinations are part of its identity. Quietly replacing an unimpressive card with a stronger one may improve the pile, but it ruins the test.
What the recreated deck needs
| Component | Quantity | Where it goes |
|---|---|---|
| Main-deck cards | 36 total: 12 from each of three houses | Shuffled together as the play deck |
| Archon identity and deck-list card | 1 | Kept outside the shuffled deck |
| Set-specific reference components | As required by that deck and set | Kept separate unless the applicable rules say otherwise |
| Backing cards and opaque sleeves | 36 matching sets | Used to make the test cards handle consistently |
The 36-card main deck can contain actions, artifacts, creatures and upgrades. On a turn, the active player chooses a house and may play, discard or use cards of that house, so a proxy must make each card's house and card type obvious at a glance.
The Archon identity card is not card number 37 in the shuffled deck. Neither are Archon Power cards or Prophecy cards. Depending on the set and deck, you may also need separate components involving Tide, token creatures or other set-specific mechanics. Keep these clearly labeled and physically separate from the main deck.
Copy the exact card, not merely the card name
A card name alone may not identify everything printed on a card in a particular deck. Enhance can place randomized bonus icons on affected cards during deck generation. Decks can also contain Mavericks, legacies or anomalies, which means that a generic image of a card may not match the version belonging to the deck you are recreating.
Audit the source deck one card at a time. For every card, verify the name, house, quantity, printed text, bonus icons and any unusual status. If two copies share a name but have different printed enhancements, treat them as two distinct files during preparation. A filename such as “01-card-name-icon” is much safer than collecting 36 vaguely labeled images in one folder.
The Archon card should also identify the recreated deck clearly. That helps prevent the list from becoming separated from the cards and makes it obvious that the deck is a casual test copy. The goal is functional accuracy, not creating something that could be confused with an official product.
Prepare readable, consistent test cards
A playtest card succeeds when the player can make decisions without stopping to decode it. Prioritize readable rules text, visible house and card-type information, recognizable art or another clear identifier, and accurate bonus icons. Tiny text and muddy images create false mistakes during testing.
Physical consistency matters too. A thinner sheet of paper mixed directly into a deck can be identified by touch or by its edge. A simple casual setup is to place each printed front into an opaque sleeve over the same kind of spare card. Use one sleeve model and one backing-card treatment across all 36 cards.
If you are preparing files for a printer, ask for its current template before arranging the final sheets. Confirm finished size, bleed, safe areas, accepted file types, color settings and whether the service expects individual fronts or imposed sheets. Our guide to custom trading-card print files explains those production terms and the checks worth making before handing off a folder.
Do not assume an MTG proxy ordering form can accept KeyForge files. ProxyFoundry's checked ordering route is designed around searching existing Magic printings or pasting a Magic deck list; it does not establish a route for creating or ordering custom KeyForge cards. A KeyForge project needs a workflow that explicitly accepts the files and configuration you have prepared.
Ghost Galaxy does publish downloadable cards for the defined Vault Assault variant, but that resource should not be treated as a general print-and-play library or authorization path for reproducing Archon decks.
Run tests that answer useful questions
Once the copy is assembled, resist the urge to declare a verdict after one unusually strong draw. KeyForge decks can present very different turns depending on house distribution, hand texture and the order in which key pieces appear.
- Play an initial series against one familiar opposing deck. Keeping the opponent stable makes the recreated deck easier to evaluate.
- Record the opening hand, any mulligan, the houses chosen each turn and cards that repeatedly remain stranded in hand.
- Note which cards improve when drawn together and which apparent combinations demand too much setup.
- Rotate through faster, slower and more disruption-heavy opposing decks after establishing the baseline.
- Swap pilots for a few games. A second player may identify house choices or sequencing lines that the first pilot overlooked.
Use the real opening procedure. The first player begins with seven cards and the second with six. Each player may take one mulligan by redrawing one fewer card, and players shuffle before offering their decks for an opponent to shuffle or cut. Avoid giving the recreated deck a convenient opening hand just because the purpose is testing.
Also test beyond the first pass through the deck. Players refill to six cards at the end of a turn, and an empty draw deck is replenished by shuffling the discard pile. A card that looks mediocre early may matter differently after the discard pile cycles back.
Do not alter the recreated list between games if the question is whether that particular deck is enjoyable or competitive in your group. Change the matchup, pilot or tactical approach instead. If you start replacing individual cards, you are no longer evaluating the known Archon deck.
Keep recreated cards in casual play
Ghost Galaxy's Tournament Rules & Guidelines version 2.7, dated May 22, 2026, says proxy decks and cards are not permitted in official KeyForge tournaments. The rules also require opaque, matching sleeves within a deck for official events. A well-made casual reproduction does not become tournament legal merely because every card is readable and sleeved. Check the current official tournament rules before attending an event.
Casual and fan-run events can use different formats or modified rules, but that does not mean proxies are automatically welcome. Ask the organizer or playgroup before arriving, explain that the deck recreates a known list and bring another legal option if the answer is uncertain. Ghost Galaxy's tournament-format guidance distinguishes official organized play from casual and fan-run activity.
If the real issue is replacing an official deck for sanctioned play, investigate the official Deck Clone program rather than relying on a homemade copy. Under Ghost Galaxy's stated limits, the Master Vault owner may use a clone, but only one instance—the original or clone—may be played at an official tournament. Review the current Deck Clone requirements for eligibility and restrictions.
The practical checklist
- Start with one known, verifiable Archon deck list.
- Reproduce exactly 36 main-deck cards: 12 from each of three houses.
- Match quantities, printed bonus icons and unusual versions instead of relying only on card names.
- Create the Archon identity and deck-list card as a separate component.
- Add only the reference cards or tokens required by that deck and set.
- Use readable fronts, identical backing cards and opaque matching sleeves.
- Keep the files organized for reprints and corrections.
- Get permission before using the recreated deck at a casual gathering.
- Use official cards or an eligible official clone for sanctioned tournament play.
Recreate accurately, then test honestly
The best KeyForge proxy deck is not a fantasy list with all the awkward cards removed. It is an accurate recreation of one Archon deck, including the strange details that make that deck worth testing.
Get the 36-card list right, keep the identity and reference components separate, preserve every printed enhancement, and make the physical cards consistent enough that they do not influence play. Then run repeated games against controlled matchups. That process will tell you something useful about the deck—without pretending a casual test copy belongs at an official tournament.