Flashcard Maker

Flashcard Maker

Generate flashcards optimized for spaced repetition systems like Anki. Good flashcards test understanding, not recall of exact wording.

Core Principles

Minimum Information Principle

Each card tests exactly ONE piece of knowledge. If a card has "and" in the answer, split it.

Active Recall Over Recognition

Cards should require the learner to produce the answer, not recognize it from options.

Context Independence

Each card must be understandable without seeing other cards. No "as mentioned in card #3".

Card Types

Type 1: Basic (Front → Back)

Best for: Definitions, facts, translations

Q: What is [term]?
A: [Concise definition in your own words]

Type 2: Cloze Deletion

Best for: Formulas, sequences, key phrases

{{c1::TCP}} operates at the {{c2::transport}} layer of the OSI model.

Type 3: Reversed

Best for: Bidirectional knowledge (term ↔ definition)

Q: What is the term for [definition]?
A: [Term]

Q: Define [term].
A: [Definition]

Type 4: Image Occlusion

Best for: Diagrams, anatomy, maps

Q: [Image with labeled part hidden]
A: [Hidden label revealed]

Type 5: Application

Best for: Deep understanding, problem-solving

Q: [Scenario or problem]
   How would you apply [concept]?
A: [Solution with reasoning]

Writing Rules

Questions

• Be specific — "What does X do?" not "Tell me about X"
• Use consistent phrasing for similar card types
• Avoid yes/no questions (too easy to guess)
• Include context cues when ambiguity exists: "In [field], what is..."

Answers

• Keep under 20 words when possible
• Use bullet points for multi-part answers (max 3 bullets)
• Bold the key term or number
• Include a mnemonic hint if the fact is hard to remember

Distractors (for multiple choice)

• Make wrong answers plausible, not obviously wrong
• Use common misconceptions as distractors
• Keep all options similar in length and style
• Never use "all of the above" or "none of the above"

Difficulty Calibration

Generate a mix: 40% easy, 40% medium, 20% hard.

Language Rules

• Match the language of the user's request by default
• If the user explicitly specifies a language (e.g., "in English", "用中文"), use that language for all output
• Keep proper nouns, brand names, and technical terms in their original language

Output Format

Standard Format (Tab-separated, Anki-importable)

Front\tBack\tTags
What is [concept]?\t[Answer]\t[topic]::[subtopic]

Rich Format (Markdown)

## Deck: [Topic Name]
Tags: #topic #subtopic

---

### Card 1
**Q:** [Question]
**A:** [Answer]
**Tags:** topic::subtopic
**Difficulty:** easy | medium | hard

---

### Card 2
...

Cloze Format

{{c1::Answer}} is the [context for the blank].
Tags: topic::subtopic

Tag Hierarchy

Use :: separator for nested tags (Anki convention):

biology::cell-biology::mitosis
programming::python::data-structures
history::world-war-2::pacific-theater

Anti-Patterns

• Verbatim textbook — copying sentences from source material (rephrase in your own words)
• Trick questions — testing gotchas instead of understanding
• Vague questions — "What about mitosis?" (what about it?)
• Essay answers — if the answer is a paragraph, the card needs splitting
• Orphan cards — facts without context (why does this matter?)
• Synonym traps — accepting only one specific word when multiple are correct

Workflow

From Text/Article

1. Identify key concepts, facts, and relationships
2. Create basic cards for terminology
3. Create application cards for concepts
4. Create comparison cards for related ideas
5. Add tags for organization
6. Review for minimum information principle

From Lecture/Video

1. Extract main claims and supporting evidence
2. Create cards for each claim-evidence pair
3. Add process/sequence cards for workflows
4. Create "why" cards for reasoning chains

Quick Checklist

1. Does each card test exactly one thing?
2. Can the card be understood without other cards?
3. Is the answer under 20 words?
4. Would a wrong answer reveal a real knowledge gap?
5. Are tags hierarchical and consistent?
6. Mix of difficulty levels present?