MCP server for MetraCards. Manage decks and flashcards from Cursor, Claude, and ChatGPT.
MetraCards MCP has 35 well-organized tools with consistent naming (verb_noun pattern) and complete input schemas. However, descriptions are uniformly terse (15-30 chars), falling well below the 50-200 char LLM-optimized baseline. Parameter descriptions exist but lack actionable detail (format constraints, ranges, dependencies). Output schemas are not documented. Error handling and recovery guidance are absent. Tool composition is sound (single responsibility, proper chaining via IDs), but the thin descriptions and missing output documentation significantly limit LLM reasoning quality.
Authenticate user via browser login with Firebase
Clear stored authentication credentials
Get current authentication status and user profile
Create a new flashcard
Create multiple flashcards in batch
Delete a flashcard
Delete multiple flashcards in batch
Generate audio for flashcard text using text-to-speech
Tool descriptions are uniformly terse (15-30 chars), far below the 50-200 char LLM-optimized baseline. Examples: 'Health check endpoint for MetraCards API' (40 chars), 'List all flashcard decks' (24 chars). Thin descriptions force LLMs to guess when/why to select tools, increasing misuse.
Output schemas are not documented. LLMs cannot plan downstream tool calls or extract required fields (e.g., deck_id from decks_create, card_id from cards_create). This forces discovery detours and breaks tool chaining.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 51 | 2026-07-28+ | v2 |
Get a specific flashcard by ID
List all flashcards, optionally filtered by deck
Update an existing flashcard
Upload media file (audio or image) for a flashcard
Create a new flashcard deck
Create multiple flashcard decks in batch
Delete a flashcard deck
Delete multiple flashcard decks in batch
Get a specific flashcard deck by ID
List all flashcard decks
Update an existing flashcard deck
Health check endpoint for MetraCards API
Create a new card review
Create multiple card reviews in batch
Delete a card review
Delete multiple card reviews in batch
Get a specific review by ID
List all card reviews
Create a new review schedule
Create multiple review schedules in batch
Delete a review schedule
Delete multiple review schedules in batch
Get a specific review schedule by ID
List all review schedules
Update an existing review schedule
Get activity statistics
Get usage statistics
Destructive operations (delete, delete_batch) lack confirmation or dry-run support. Agents can irreversibly delete decks, cards, reviews, or schedules without user confirmation, risking data loss.
Parameter descriptions lack actionable constraints. Example: 'Optional timestamp to fetch decks updated since this time' does not specify format (ISO 8601? Unix epoch?). 'Optional deck color' does not specify valid values (hex? named colors?). LLMs cannot validate inputs without explicit format/range guidance.
No error handling or recovery guidance visible in tool definitions. LLMs receive raw API errors with no actionable next steps (e.g., 'User not found. Try auth_login() first.'). This blocks agent self-correction.