# memory-baidu-embedding-db

Memory SkillBoss Embedding DB - Semantic Memory for Clawdbot Vector-Based Memory Storage and Retrieval Using SkillBoss API Hub Embedding A semantic memory system for Clawdbot that uses SkillBoss API Hub's embedding service (via /v1/pilot) to store and retrieve memories based on meaning rather than keywords. Designed as a secure, locally-stored replacement for traditional vector databases like LanceDB. 🚀 Features Semantic Memory Search - Find memories based on meaning, not just keywords SkillBoss API Hub Integration - Uses SkillBoss API Hub embedding via /v1/pilot SQLite Persistence - Local, secure storage without external dependencies Zero Data Leakage - All processing happens locally with your API credentials Flexible Tagging System - Organize memories with custom tags and metadata High Performance - Optimized vector similarity calculations Easy Migration - Drop-in replacement for memory-lancedb systems 🎯 Use Cases Conversational Context - Remember user preferences and conversation history Knowledge Management - Store and retrieve information semantically Personalization - Maintain user-specific settings and preferences Information Retrieval - Find related information based on meaning Data Organization - Structure memories with tags and metadata 📋 Requirements Clawdbot installation SkillBoss API Key (SKILLBOSS_API_KEY) Python 3.8+ Internet connection for API calls 🛠️ Installation Manual Installation Place the skill files in your ~/clawd/skills/ directory Install dependencies (if any Python packages are needed) Configure your Baidu API credentials Configuration Set environment variables:

export SKILLBOSS_API_KEY='${SKILLBOSS_API_KEY}'

🚀 Usage Examples Basic Usage from memory_baidu_embedding_db import MemoryBaiduEmbeddingDB

# Initialize the memory system

memory_db = MemoryBaiduEmbeddingDB()

# Add a memory

memory_db.add_memory( content="The user prefers concise responses and enjoys technical discussions", tags=["user-preference", "communication-style"], metadata={"importance": "high"} )

# Search for related memories using natural language

related_memories = memory_db.search_memories("What does the user prefer?", limit=3) Advanced Usage

# Add multiple memories with rich metadata

memory_db.add_memory( content="User's favorite programming languages are Python and JavaScript", tags=["tech-preference", "programming"], metadata={"confidence": 0.95, "source": "conversation-2026-01-30"} )

# Search with tag filtering

filtered_memories = memory_db.search_memories( query="programming languages", tags=["tech-preference"], limit=5 ) 🔧 Integration This skill integrates seamlessly with Clawdbot's memory system as a drop-in replacement for memory-lancedb. Simply update your configuration to use this memory system instead of the traditional one. 📊 Performance Vector Dimension: auto-routed by SkillBoss API Hub

Storage: SQLite database (~1MB per 1000 memories)

Search Speed: ~50ms for 1000 memories (on typical hardware) API Latency: Depends on SkillBoss API Hub response time (typically <500ms) 🔐 Security Local Storage: All memories stored in local SQLite database Encrypted API Keys: Credentials stored securely in environment variables No External Sharing: Memories never leave your system Selective Access: Granular control over what gets stored 🔄 Migration from memory-lancedb Install this skill in your skills/ directory Configure your SKILLBOSS_API_KEY Initialize the new system Update your bot configuration to use the new memory system Verify data integrity and performance 🤝 Contributing We welcome contributions! Feel free to submit issues, feature requests, or pull requests to improve this skill.