# MTG Bench (remote · mtgbench.com)

Magic: The Gathering card search, rules, rulings, deck analysis, brackets, and combo detection.

- Trust score: 68/100 (medium)
- Change this week: +3
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-08-19

## Components

- remote · `mtgbench.com`: 68/100 (this document), [markdown](https://verifymcp.io/servers/com-mtgbench-mtgbench/mtgbench.md), [page](https://verifymcp.io/servers/com-mtgbench-mtgbench/mtgbench)

## Channel facts

- Endpoint: `https://mtgbench.com/mcp`
- Transports: `streamable-http`
- Auth: `none`
- Version: `0.52.2`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically against the live server, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-08-19.

- **Endpoint Security**: 63/100
  - The endpoint's TLS certificate is valid, in date, and uses a strong key.
  - Authorisation not fully verified: no authorisation is required to call this server, and 19 tool(s) never declared a destructiveHint. The MCP spec treats an absent hint as destructive by default, so we cannot call this surface safe.
  - HTTPS is enforced; there's no plaintext access path.
  - The HSTS (Strict-Transport-Security) header is present.
  - DNSSEC check failed: this domain isn't protected by DNSSEC.
- **Transport & Reachability**: 100/100
  - Verified streamable-http transport via a live MCP handshake.
- **Schema Quality & AI Usability**: 79/100
  - 100% of prompts and resources have a non-trivial description (not blank, and not just the item's name).
  - AI-judged instruction clarity (good).
  - Context-footprint check failed: tool/resource definitions use about 2462 tokens (~111/item across 22 items; 19 tools + 3 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 33/100
  - Stability observed for 10 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 71/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 0% of tool parameters carry a description.
  - Structured output schemas are declared (100% of tools); any adoption earns full credit.
- **Capabilities**: 100/100
  - Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.
  - Supports UI / widget rendering.

## Install

### Claude

```bash
claude mcp add --transport http com-mtgbench-mtgbench https://mtgbench.com/mcp
```

### Codex

```toml
[mcp_servers.com-mtgbench-mtgbench]
url = "https://mtgbench.com/mcp"
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "com-mtgbench-mtgbench": {
      "type": "remote",
      "url": "https://mtgbench.com/mcp",
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add com-mtgbench-mtgbench --url https://mtgbench.com/mcp --transport streamable-http
```

### Hermes

```yaml
mcp_servers:
  com-mtgbench-mtgbench:
    url: "https://mtgbench.com/mcp"
```

### Other

```json
{
  "mcpServers": {
    "com-mtgbench-mtgbench": {
      "type": "http",
      "url": "https://mtgbench.com/mcp"
    }
  }
}
```

The mcpServers block is a cross-client convention. Remote transports vary, so check your client's docs.

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-08-18 (score 68, +1)

No change was recorded against any check on this day. Stability & Change Management went from 27 to 30. That category is still filling its 30-day observation window: 8 days of observed history at the previous scan, 9 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-08-16 (score 67, +1)

No change was recorded against any check on this day. Stability & Change Management went from 20 to 23. That category is still filling its 30-day observation window: 6 days of observed history at the previous scan, 7 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-08-14 (score 66, +1)

No change was recorded against any check on this day. Stability & Change Management went from 13 to 17. That category is still filling its 30-day observation window: 4 days of observed history at the previous scan, 5 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-08-12 (score 65, +1)

- [functional] Server version: v0.53.0 → dev

### 2026-08-11 (score 64, −1)

- [functional] We updated how we score, so this day's move reflects our rubric, not a change to the server

### 2026-08-10 (score 65, 0)

- [functional improvement] Stability: unverified → 0.03

### 2026-08-09 (score 65)

First indexed and scored.

## MCP tools (19)

### `check_deck_legality` (~142 tokens)

Validate a decklist against a constructed format's rules (banned cards, deck size, copy limits, and in Commander the color-identity rule). Use this rather than recalling ban lists from memory: bans changed four times in 2026 alone, and a model's training cutoff makes stale ban data one of the most common wrong answers in Magic. Pass `commander` for Commander decks: it counts toward the 100, and it is what makes the color-identity check possible (without it the response says color identity went unchecked).

Input parameters:

- `cards` (null|array, required)
- `commander` (string)
- `format` (string, required)

Output parameters:

- `legal` (boolean)
- `notes` (null|array)
- `violations` (null|array)

### `commander_staples` (~151 tokens)

The cards most commonly played with a given commander, from EDHREC's aggregate inclusion data: 691,958 commander-card pairings across 2,631 commanders. Returns each card's deck count, the share of that commander's decks running it, and its synergy score. Synergy matters as much as popularity: a high deck count with LOW synergy is a card everyone plays anyway (Sol Ring), while high synergy is a card that is unusually good with this commander specifically. Lead with the latter when recommending. Use this to ground deckbuilding in what people actually play instead of guessing.

Input parameters:

- `commander` (string, required)
- `limit` (integer)
- `minSynergy` (number)

Output parameters:

- `cards` (null|array)
- `commander` (string)
- `totalDecks` (integer)

### `deck_stats` (~122 tokens)

Deterministic statistical analysis of a decklist (any format): mana curve with creature split, card types, color pips vs mana sources, functional buckets (ramp/draw/removal/wipes/counters/recursion/protection/tutors + bracket signals) with typical Commander bands, opening-hand probabilities (hypergeometric), a suggested land count, meta brew score, deck price, tribal and keyword counts. Pure data — no judgment calls.

Input parameters:

- `cards` (null|array, required)
- `commander` (string)
- `format` (string)

Output parameters:

- `avgManaValue` (number)
- `bandBracket` (integer)
- `brew` (null|object)
- `buckets` (object)
- `curve` (null|array)
- `dorkSources` (integer)
- `identity` (null|object)
- `landCount` (integer)
- `landSources` (integer)
- `medianManaValue` (number)
- `nonlandCount` (integer)
- `notes` (null|array)
- `openingHand` (null|array)
- `pips` (object)
- `price` (null|object)
- `rockSources` (integer)
- `sources` (object)
- `suggestedLands` (integer)
- `targetSize` (integer)
- `topKeywords` (null|array)
- `topSubtypes` (null|array)
- `totalCards` (integer)
- `types` (object)
- `unknownCards` (null|array)

### `find_combos` (~121 tokens)

Detect known combos in a decklist using Commander Spellbook's curated database — combos fully present, plus near-misses one card away (with that card's price). Pass `format` (and `commander` for Commander) so near-miss suggestions are filtered to legal, in-color-identity adds. Use this instead of recalling combos from memory; also feeds rate_deck's compact-combo check.

Input parameters:

- `cards` (null|array, required)
- `commander` (string)
- `format` (string)
- `limit` (integer)

Output parameters:

- `combos` (null|array)
- `nearMisses` (null|array)
- `notes` (null|array)

### `find_synergies` (~142 tokens)

Given one or more cards the user ALREADY has, find cards that work with them. Seeded by example rather than by description: pass the cards, get partners. Use semantic_card_search instead to describe a desired effect in words, and search_cards for structured filters. `limit` defaults to 10 and caps at 100 per call.

Input parameters:

- `cardNames` (null|array, required)
- `colors` (null|array)
- `format` (string)
- `limit` (integer)
- `manaValueMax` (null|number)
- `manaValueMin` (null|number)
- `types` (null|array)

Output parameters:

- `results` (null|array)

### `get_card` (~87 tokens)

Fetch full detail for ONE card by exact name or id: oracle text, mana cost, type line, format legalities, and a brief summary of its printings. Use get_card_printings instead when the complete printing list matters (choosing a specific art, set, or finish).

Input parameters:

- `id` (string)
- `name` (string)
- `slug` (string)

Output parameters:

- `card` (object)

### `get_card_image` (~117 tokens)

Fetch a card's image. The returned image content block IS the display — the client renders it directly; do not build custom HTML/widgets around the imageUrl, since sandboxed widget environments block external image hosts. Optionally pick a specific printing (setCode) or size (small, normal, large, png, art_crop, border_crop, or WEBP: thumb, grid, display).

Input parameters:

- `id` (string)
- `name` (string)
- `setCode` (string)
- `version` (string)

Output parameters:

- `imageUrl` (string)
- `name` (string)
- `setCode` (string)

### `get_card_printings` (~72 tokens)

List EVERY printing of a card (set, rarity, artist, finish, collector number). get_card already returns a short printings summary, so reach for this only when the full list is the point.

Input parameters:

- `id` (string)
- `name` (string)
- `slug` (string)

Output parameters:

- `printings` (null|array)

### `get_card_rulings` (~100 tokens)

Fetch the OFFICIAL Wizards rulings for one card, with their dates. These are Wizards' own answers to the corner cases players argue about. Use this for any "how does this card actually work" question, and quote the ruling rather than paraphrasing from memory — rulings are among the most frequently hallucinated things in Magic. Use search_rules instead for general mechanics not tied to a specific card.

Input parameters:

- `name` (string, required)

Output parameters:

- `card` (string)
- `count` (integer)
- `rulings` (null|array)

### `get_rule` (~63 tokens)

Fetch a Comprehensive Rules entry and its lettered subrules by exact number (e.g. "601.2" returns 601.2 and 601.2a-i). Use after search_rules to read a rule's full context.

Input parameters:

- `ref` (string, required)

Output parameters:

- `rules` (null|array)

### `my_collection` (~88 tokens)

Summarize the user's imported collection: how many distinct cards, total copies, and where it was imported from. This is the SUMMARY only. Use search_my_collection to look up individual owned cards, and what_can_i_build to rank commanders they could assemble. Requires a free API key (send it as "X-API-Key: <key>" or "Authorization: Bearer <key>").

Output parameters:

- `distinctCards` (integer)
- `source` (string)
- `totalCopies` (integer)

### `rate_deck` (~111 tokens)

Assess a Commander deck's official WotC bracket (1-5). Returns deterministic signals (Game Changers from WotC's curated list, mass land denial, tutors, fast mana, free interaction), a bracket FLOOR from the official decision tree, and the rubric — then YOU finish the assessment: check for compact two-card win combos and weigh soft signals per the assessmentGuidance, and state the final bracket with evidence.

Input parameters:

- `cards` (null|array, required)
- `commander` (string)

Output parameters:

- `assessmentGuidance` (string)
- `bracketFloor` (integer)
- `bracketFloorReason` (string)
- `combosPresent` (null|array)
- `deck` (object)
- `earlyCombo` (string)
- `notes` (null|array)
- `rubric` (string)
- `signals` (object)
- `unknownCards` (null|array)

### `resolve_card_names` (~128 tokens)

Check whether card names are REAL, in bulk. Returns exact / not_found per name, with the canonical spelling and id for the ones that exist. Call this reflexively before presenting any decklist or card recommendation you generated: the most common failure in Magic assistance is confidently naming cards that do not exist, or that exist with different text. Cheap, batched, and safe to call on every card you are about to mention. Accepts either a `names` array or a pasted `text` decklist.

Input parameters:

- `names` (null|array)
- `text` (string)

Output parameters:

- `found` (integer)
- `notFound` (integer)
- `results` (null|array)

### `search_cards` (~231 tokens)

Find cards by NAME or by structured filters (colors, color identity, types, mana value, format legality, sets). Use when the query can be expressed as filters: "red instants under 3 mana", "legendary artifacts in Standard", "cards named Bolt". Use semantic_card_search instead when the user describes an EFFECT rather than a filter ("cards that make a treasure when I attack"). Use find_synergies instead when they name cards they already have and want things that pair with them. Searches every Magic card; use search_my_collection to search only what the user owns. `limit` defaults to 25 and caps at 100 per call.

Input parameters:

- `colorIdentity` (null|array)
- `colors` (null|array)
- `format` (string)
- `limit` (integer)
- `manaValueMax` (null|number)
- `manaValueMin` (null|number)
- `name` (string)
- `sets` (null|array)
- `supertypes` (null|array)
- `types` (null|array)

Output parameters:

- `cards` (null|array)

### `search_my_collection` (~158 tokens)

Search only the cards the user OWNS, by name, colors, or type. Set notInDecks to surface owned cards not currently used in any of their decks (the "what is sitting idle in my binder" question). Use search_cards instead to search every Magic card regardless of ownership. Returns at most 100 owned cards per call, with the full match count in `total`. Requires a free API key (send it as "X-API-Key: <key>" or "Authorization: Bearer <key>").

Input parameters:

- `colors` (null|array)
- `limit` (integer)
- `notInDecks` (boolean)
- `query` (string)
- `types` (null|array)

Output parameters:

- `cards` (null|array)
- `total` (integer)

### `search_rules` (~100 tokens)

Semantic search over the Magic Comprehensive Rules, glossary, and 84k card-specific rulings. Use for any rules/interaction question; answer FROM the returned chunks and cite rule numbers (e.g. 601.2a). Filter kinds to ["ruling"] for card-specific questions or ["rule","glossary"] for general mechanics.

Input parameters:

- `kinds` (null|array)
- `limit` (integer)
- `query` (string, required)

Output parameters:

- `results` (null|array)

### `semantic_card_search` (~176 tokens)

Vector search over card TEXT, for describing an effect that cannot be written as a filter: "cards that make a token when I attack", "punish opponents for drawing", "does what Rhystic Study does but cheaper". This is a query Scryfall syntax cannot express. Use search_cards instead for structured filters (color, type, mana value, set). Use find_synergies instead when the user names specific cards and wants partners for them. `limit` defaults to 10 and caps at 100 per call.

Input parameters:

- `colors` (null|array)
- `format` (string)
- `limit` (integer)
- `manaValueMax` (null|number)
- `manaValueMin` (null|number)
- `query` (string, required)
- `types` (null|array)

Output parameters:

- `results` (null|array)

### `upgrade_deck` (~168 tokens)

Suggest priced deck upgrades: cards to add (within an optional budget) and cards to cut. TWO-PHASE: first call WITHOUT `strategy` to get a deck analysis and instructions for deriving the deck's strategy; then call again WITH your derived `strategy` to get recommendations. Cuts are theme-alignment suggestions relative to the stated strategy, not card-quality judgments. Prices are Card Kingdom/TCGplayer retail USD.

Input parameters:

- `budget` (null|number)
- `cards` (null|array, required)
- `commander` (string)
- `format` (string, required)
- `maxAdds` (integer)
- `maxCardPrice` (null|number)
- `maxCuts` (integer)
- `ownedOnly` (boolean)
- `strategy` (string)

Output parameters:

- `adds` (null|array)
- `budgetRemaining` (null|number)
- `collectionTip` (null|object)
- `cuts` (null|array)
- `deck` (null|object)
- `legalityIssues` (null|array)
- `notes` (null|array)
- `phase` (string)
- `strategy` (string)
- `strategyPrompt` (string)
- `totalCost` (null|number)
- `unknownCards` (null|array)

### `what_can_i_build` (~122 tokens)

Given the user's imported card collection, rank the Commander decks they can most nearly build right now, with the most-played cards they are still missing. Answers "what should I sleeve up tonight" from cards they physically have. Requires a free API key (send it as "X-API-Key: <key>" or "Authorization: Bearer <key>") so the server knows whose collection to read; one is issued instantly at https://mtgbench.com/account/api-keys.

Input parameters:

- `limit` (integer)
- `missing` (integer)

Output parameters:

- `commanders` (null|array)
- `note` (string)

## Diagnostics

Captured diagnostic sections: TLS, DNSSEC, Authorisation, Transports. The full working is on the page: https://verifymcp.io/servers/com-mtgbench-mtgbench/mtgbench#diagnostics

## Score history

- 2026-08-19: 68
- 2026-08-18: 68
- 2026-08-17: 67
- 2026-08-16: 67
- 2026-08-15: 66
- 2026-08-14: 66
- 2026-08-13: 65
- 2026-08-12: 65
- 2026-08-11: 64
- 2026-08-10: 65
- 2026-08-09: 65

## Links

- Remote endpoint: https://mtgbench.com/mcp
- Website: https://mtgbench.com/
- Changelog RSS feed: https://verifymcp.io/servers/com-mtgbench-mtgbench/mtgbench.xml
- Changelog JSON feed: https://verifymcp.io/servers/com-mtgbench-mtgbench/mtgbench.json
- HTML version of this page: https://verifymcp.io/servers/com-mtgbench-mtgbench/mtgbench