Short answers to the product, storage, and AI workflow questions that come up most.
Adria MTG is a Magic: The Gathering toolkit for Limited drafting, Commander deckbuilding, and Standard metagame exploration. The app combines deterministic game logic, external card/meta data, and AI assistance where it adds value.
Yes. Starting a draft requires an account. You can browse the Standard metagame and explore commander tools without signing up, but the draft simulator and draft history require you to be logged in. Creating an account is free.
The product is currently free to use during open beta. Future paid tiers may add premium features, but the current core workflows are available without a subscription. See Pricing for the current hosted coaching and BYO key model.
The draft coach can provide post-pick analysis, pick grading, and pre-pick reasoning prompts. The scoring logic also uses draft context, curve pressure, color commitment, and 17Lands-backed card performance where available.
Yes. Coaching is optional. You can draft with the coach off and still use the simulator and post-draft review flows.
The app supports OpenRouter and Anthropic. You can store an OpenRouter key locally in Settings (browser only) or, if logged in, save an encrypted server-side key on your Profile, which keeps the raw key off the client after save.
There are two paths today. The Settings page stores an OpenRouter key in browser localStorage for local-only use. Logged-in users can also save a server-side encrypted key on their Profile, which supports provider selection and keeps the raw key off the client after save.
A whistle is consumed by two AI coaching modes: the Thinking Guide(pre-pick questions that walk you through the key considerations for a pack — curve, color signals, BREAD, Quadrant Theory — and always end with a direct pick recommendation so you're never left stuck) and the Analysis (post-pick AI explanation of why your pick was strong or weak). Every pick always shows a rule-based Quick Score— a letter grade (A+ to F) based on card power, color fit, and draft context — which never costs a whistle. When a free account's 7 whistles are spent, the Quick Score remains visible on every remaining pick but the AI-powered Analysis and Thinking Guide are unavailable until the next draft. Free users get 7 whistles per draft, Pro removes that cap, and BYO key flows route the model cost to your own provider instead. The Pricing and Billing pages explain the current usage model.
The app focuses on Standard-legal sets and keeps the set list in sync from external card data sources. Draft packs use real card data, and the app layers in cached ratings and heuristics on top of that pool.
Pick grades measure decision quality relative to your options, not a card’s absolute power level. The engine scores every card in the pack for your specific draft state, then grades your pick against those alternatives: taking the best available card earns an A+; taking the worst earns an F. A high grade means you made the right call given what was in front of you — it does not mean the card is objectively strong. A mediocre card can earn an A+ from a weak pack. When the AI coach assigns a high grade and the pack was limited, it will say so in the commentary.
Yes. Adria uses 17Lands-backed statistics where available to enrich card evaluation and draft feedback. That data complements, rather than replaces, the app's own draft-state logic.
Yes. Draft sessions are persisted to your account and can be resumed at any time from the draft hub.
Use the Import Guide. It covers supported export styles, Limited versus Standard versus Commander imports, and what happens after Adria parses your list.
Commander support includes commander search, AI-assisted brewing, deck import, saved deck editing, color-identity validation, bracket estimation, and deck analysis on saved lists.
Yes. The app tracks bracket context, checks color identity, and flags game-changer style cards relative to the selected bracket.
Standard metagame views are built from fetched external meta data, normalized into archetypes, matchup views, and AI-assisted analysis. The codebase currently pulls that data from MTGGoldfish-oriented fetchers and related parsers.
It is the part of the app that tries to identify underplayed or exploitable pockets in the current Standard field and turn them into deck ideas or brew directions.
The app relies on external services for card data and images (Scryfall), draft-performance stats (17Lands), LLM access (OpenRouter and Anthropic), and Standard metagame sourcing.