Exploring fun, creative, and human-friendly ways to memorize high-entropy data like passwords and crypto keys 🎨🎵🎮
Easy Keys is a research project investigating innovative approaches to help humans memorize high-entropy data (like cryptocurrency keys) in ways that are:
- Fun & Engaging: Turn security into games, stories, and experiences
- Multi-Sensory: Beyond just words - colors, sounds, movements, and more
- Personally Meaningful: Adapt to individual learning styles
- Scientifically Grounded: Based on cognitive science and memory research
Our initial 4-word mnemonic system demonstrates a critical insight: the first word is absolutely essential. Without the correct first word ("public" or "secret"), the remaining 3 words become completely meaningless.
✅ Valid: public mountain river storm
❌ Invalid: secret mountain river storm (wrong first word)
❌ Invalid: mountain river storm (missing first word = meaningless)
We've implemented multiple innovative encoding approaches:
- Color Grids: 4x4 grids of perceptually distinct colors
- Geometric Patterns: Shapes, rotations, and spatial arrangements
- Visual Stories: Comic-strip style narrative panels
- Melodies: Notes, octaves, and rhythms you can hum
- Soundscapes: 3D environmental audio experiences
- Chord Progressions: Emotional journeys through harmony
- Memory Palaces: Virtual rooms with memorable objects
- Dance Sequences: Body movements and gestures
- Map Journeys: Real-world landmark navigation
- RPG Quests: Heroes, battles, and epic rewards
- Puzzle Games: Levels, challenges, and achievements
- Leaderboards: Competitive stats and memorable plays
- Personalized Learning: Adapts to your cognitive style
- Synesthetic Encoding: Cross-sensory experiences
- Adaptive Systems: Learns from your performance
- Memory is Multi-Dimensional: Different people remember differently
- Stories Beat Random: Narrative structures improve retention 10-100x
- Fun Improves Security: Engaged users create stronger, more memorable keys
- Multiple Pathways Help: Redundancy through different senses aids recall
- Words Alone: Limited to ~11 bits per word, requiring many words for full entropy
- Cultural Barriers: Word meanings vary across languages and cultures
- Cognitive Load: 14+ words becomes challenging for most users
- Static Nature: Words don't adapt to individual preferences
We haven't yet explored but are excited about:
- Voice Recognition: Unique vocal patterns and speech characteristics
- Video Entropy: Facial expressions, gestures, and movement patterns
- Biometric Integration: Combining memorable patterns with personal traits
- AR/VR Experiences: Immersive 3D memory environments
- Social Proof Systems: Distributed verification without central authority
- Original 4-word system: Proves concept but limited coverage (~13% of address)
- Perfect fidelity: 17 words required for 100% address representation
- Optimized system: 14 words achieved (17.6% reduction)
- Theoretical limit: 9-10 words identified (requires advanced encoding)
- Alternative modalities: Visual/audio/spatial can encode 20-50 bits per element
See RESEARCH.md for optimization strategies and CREATIVE_ALTERNATIVES.md for non-word approaches.
# Install UV package manager (if not already installed)
curl -LsSf https://astral.sh/uv/install.sh | sh
# Clone and run
git clone https://github.com/dirvine/easy_keys.git
cd easy_keys
uv run python -m easy_keys --help# Test all creative encoding systems
uv run python test_creative_encoders.py
# Or try individual encoders:
uv run python -m easy_keys.visual_encoder # Color grids and patterns
uv run python -m easy_keys.musical_encoder # Melodies and soundscapes
uv run python -m easy_keys.spatial_encoder # Memory palaces and movements
uv run python -m easy_keys.gamified_encoder # RPG quests and puzzles
uv run python -m easy_keys.multimodal_encoder # Personalized combinations# Encode Ethereum address as public
uv run python -m easy_keys encode --type public --address 0x742d35Cc6cA6CE02036dBdE3e8Bf13f96Db9F141
# Encode Bitcoin address as secret
uv run python -m easy_keys encode --type secret --address 1A1zP1eP5QGefi2DMPTfTL5SLmv7DivfNa
# Validate a mnemonic
uv run python -m easy_keys validate public mountain river storm
# Demonstrate first word criticality
uv run python -m easy_keys criticality mountain river storm
# Run system tests
uv run python -m easy_keys test --verbose🎨 VISUAL: 4x4 Color Grid
[🟥][🟦][🟩][🟨] "In the north: sunset, ocean, forest, sunshine"
[🟪][🟦][🟧][🟩] "In the east: twilight, ocean, flame, forest"
[🟦][🌸][🟩][🟩] "In the south: ocean, blossom, forest, forest"
[🟧][🟦][🟦][🟥] "In the west: flame, ocean, ocean, sunset"
🎵 MUSICAL: Melody Sequence
Notes: C4 D5 E4 F5 G4 A5 B4 C5
Rhythm: ♩ ♪♪ ♩ ♬♬♬♬
Story: "ascending like sunrise → bouncing like a ball → wandering freely"
🏛️ SPATIAL: Memory Palace Journey
"In the kitchen, you see a large black dishwasher placed left.
Leading to the garage with its tiny white bikes by the window..."
🎮 GAMIFIED: RPG Quest
"You are a Level 15 Rogue with earthquake technique
Journey to the Underwater Palace, defeat the Giant Spider
Face the Ancient Goblin - weakness: stealth!"
🌐 MULTI-MODAL: Personal Experience
"In a beach, discover a circle glowing with orange light.
Hear syncopated violin filling you with mystery.
The soft surface tastes bitter, guiding you west..."
- Address → Integer: Convert crypto address to numerical value
- Integer → Words: Map to 3 words from 16K+ dictionary (14 bits each)
- Checksum Calculation:
checksum = (word2_index + word3_index + word4_index) % 2 - Parity Rule:
- Even checksum (0) → requires "public" first word
- Odd checksum (1) → requires "secret" first word
- Critical Dependency: Wrong first word = invalid mnemonic
# 1. Encode
address = "0x742d35Cc6cA6CE02036dBdE3e8Bf13f96Db9F141"
mnemonic = encode_address(address, "public")
# → ["public", "mountain", "river", "storm"]
# 2. Validate checksum
checksum = calculate_checksum(["mountain", "river", "storm"]) # → 0 (even)
# Even checksum requires "public" first word ✅
# 3. Test wrong first word
wrong_mnemonic = ["secret", "mountain", "river", "storm"]
# Checksum is still 0 (even) but "secret" requires odd → INVALID ❌
# 4. Decode
key_type, addr_int = decode_mnemonic(["public", "mountain", "river", "storm"])
# → ("public", 1234567890...)- Ethereum (ETH): Standard addresses (0x...)
- Bitcoin (BTC): Legacy (1...), SegWit (3...), Bech32 (bc1...)
uv run python -m easy_keys encode --type public --address 0x742d35Cc...uv run python -m easy_keys decode public mountain river stormuv run python -m easy_keys validate public mountain river storm --detaileduv run python -m easy_keys criticality mountain river storm
# Shows which first word is required and why others failecho "0x742d35Cc6cA6CE02036dBdE3e8Bf13f96Db9F141" > addresses.txt
uv run python -m easy_keys batch --input-file addresses.txt --output-file results.txtuv run python -m easy_keys test --verboseuv run python -m easy_keys examplesThe system includes built-in tests that prove the first word is critical:
# Test that demonstrates the critical dependency
uv run python -m easy_keys criticality ability absence academy
# Output shows:
# ✅ Required first word: secret
# ✅ Checksum: 1 (odd)
# 🧪 Testing all possible first words:
# ✅ 'secret' + words → VALID
# ❌ 'public' + words → INVALID (Checksum mismatch)
# ❌ 'invalid' + words → INVALID (must be 'public' or 'secret')🚨 EDUCATIONAL PURPOSE ONLY 🚨
- Never use with real private keys in production
- This is a simplified encoding scheme, not cryptographically secure
- Real applications should use established standards like BIP39
- Always validate inputs and sanitize outputs
- No network transmission of sensitive data
Run the comprehensive test suite:
# Unit tests
uv run python -m pytest tests/ -v
# Integration tests with coverage
uv run python -m pytest tests/ --cov=easy_keys --cov-report=html
# Built-in system tests
uv run python -m easy_keys test --verboseeasy_keys/
├── easy_keys/ # Main package
│ ├── cli.py # Command-line interface
│ ├── encoder.py # Core encoding logic
│ ├── validator.py # Validation algorithms
│ ├── dictionary.py # Word dictionary management
│ ├── crypto_utils.py # Address utilities
│ ├── perfect_encoder.py # 17-word perfect fidelity
│ ├── optimal_encoder.py # 14-word optimized system
│ ├── hybrid_encoder.py # Hybrid optimization attempts
│ ├── optimized_encoder.py # Optimization strategies
│ ├── radical_optimizer.py # Advanced compression research
│ └── theoretical_limits.py # Mathematical limits analysis
├── tests/ # Comprehensive tests
├── data/ # Word dictionary files
├── examples/ # Usage examples
├── RESEARCH.md # Optimization research
└── SPECIFICATION.md # Technical specification
- Python 3.11+
- UV package manager
- Dependencies:
cryptography,base58,click
- Encoding speed: 1000+ addresses/second
- Dictionary size: 2,048 words (11 bits each)
- Address coverage: 33 bits (~13% of address)
- Memory usage: Minimal (<1MB)
- Perfect fidelity: 17 words (100% address coverage)
- Optimized system: 14 words (17.6% reduction)
- Theoretical minimum: 9-10 words (identified)
- All systems maintain: Critical first word property
# Ethereum examples
uv run python -m easy_keys encode --type public --address 0xd8dA6BF26964aF9D7eEd9e03E53415D37aA96045
uv run python -m easy_keys encode --type secret --address 0x742d35Cc6cA6CE02036dBdE3e8Bf13f96Db9F141
# Bitcoin examples
uv run python -m easy_keys encode --type public --address 1A1zP1eP5QGefi2DMPTfTL5SLmv7DivfNa
uv run python -m easy_keys encode --type secret --address bc1qw508d6qejxtdg4y5r3zarvary0c5xw7kv8f3t4- Fork the repository
- Create a feature branch
- Add tests for new functionality
- Ensure all tests pass:
uv run python -m pytest - Submit a pull request
MIT License - see LICENSE file for details.
- Repository: https://github.com/dirvine/easy_keys
- Issues: https://github.com/dirvine/easy_keys/issues
- UV Package Manager: https://github.com/astral-sh/uv
This project explores:
- Cognitive Science: How humans best memorize high-entropy data
- Multi-Modal Learning: Leveraging different sensory pathways
- Gamification: Making sec 6874 urity fun and engaging
- Decentralized Systems: Human-memorable keys without central authorities
- Future Interfaces: Voice, video, AR/VR for key management
Perfect for research in cryptography, HCI, cognitive science, and security UX.
We believe the future of secure, decentralized systems depends on making them genuinely human-friendly. This means:
- Moving beyond words to embrace how humans naturally remember
- Making security joyful rather than a chore
- Personalizing to individual strengths and preferences
- Building resilience through multiple memory pathways
- Keeping it decentralized - your keys, your control
- 🎤 Voice Integration: Unique vocal patterns for key generation
- 📹 Video Entropy: Facial expressions and gesture recognition
- 🥽 AR/VR Experiences: Immersive memory palaces
- 🧬 Biometric Fusion: Combining memorable with measurable
- 🌐 Social Verification: Distributed trust without centralization
Security doesn't have to be boring. By combining art, music, games, and science, we can create systems that are both mathematically secure and genuinely delightful to use.
The journey from 17 words to 14 was just the beginning. The real adventure is discovering how many different ways humans can reliably remember high-entropy data - and having fun while doing it!
Remember: The first word is critical... but the future might be a color, a melody, or even a dance! 🎨🎵🕺