npx skills add https://github.com/chrislemke/stoffy --skill 'AgentDB Vector Search'利用 AgentDB 的高性能向量数据库实现基于向量的语义搜索,其操作速度比传统解决方案快 150 倍至 12,500 倍。具备 HNSW 索引、量化和亚毫秒级搜索(<100µs)功能。
# 使用默认维度初始化(OpenAI ada-002 为 1536)
npx agentdb@latest init ./vectors.db
# 为不同的嵌入模型自定义维度
npx agentdb@latest init ./vectors.db --dimension 768 # sentence-transformers
npx agentdb@latest init ./vectors.db --dimension 384 # all-MiniLM-L6-v2
# 使用预设配置
npx agentdb@latest init ./vectors.db --preset small # <10K 向量
npx agentdb@latest init ./vectors.db --preset medium # 10K-100K 向量
npx agentdb@latest init ./vectors.db --preset large # >100K 向量
# 用于测试的内存数据库
npx agentdb@latest init ./vectors.db --in-memory
# 基础相似性搜索
npx agentdb@latest query ./vectors.db "[0.1,0.2,0.3,...]"
# Top-k 结果
npx agentdb@latest query ./vectors.db "[0.1,0.2,0.3]" -k 10
# 带相似度阈值(余弦相似度)
npx agentdb@latest query ./vectors.db "0.1 0.2 0.3" -t 0.75 -m cosine
# 不同的距离度量
npx agentdb@latest query ./vectors.db "[...]" -m euclidean # L2 距离
npx agentdb@latest query ./vectors.db "[...]" -m dot # 点积
# 用于自动化的 JSON 输出
npx agentdb@latest query ./vectors.db "[...]" -f json -k 5
# 带距离的详细输出
npx agentdb@latest query ./vectors.db "[...]" -v
# 将向量导出为 JSON
npx agentdb@latest export ./vectors.db ./backup.json
# 从 JSON 导入向量
npx agentdb@latest import ./backup.json
# 获取数据库统计信息
npx agentdb@latest stats ./vectors.db
import { createAgentDBAdapter, computeEmbedding } from 'agentic-flow/reasoningbank';
// 使用向量搜索优化进行初始化
const adapter = await createAgentDBAdapter({
dbPath: '.agentdb/vectors.db',
enableLearning: false, // 仅向量搜索
enableReasoning: true, // 启用语义匹配
quantizationType: 'binary', // 内存减少 32 倍
cacheSize: 1000, // 快速检索
});
// 存储带嵌入的文档
const text = "The quantum computer achieved 100 qubits";
const embedding = await computeEmbedding(text);
await adapter.insertPattern({
id: '',
type: 'document',
domain: 'technology',
pattern_data: JSON.stringify({
embedding,
text,
metadata: { category: "quantum", date: "2025-01-15" }
}),
confidence: 1.0,
usage_count: 0,
success_count: 0,
created_at: Date.now(),
last_used: Date.now(),
});
// 使用 MMR(最大边际相关性)进行语义搜索
const queryEmbedding = await computeEmbedding("quantum computing advances");
const results = await adapter.retrieveWithReasoning(queryEmbedding, {
domain: 'technology',
k: 10,
useMMR: true, // 多样化结果
synthesizeContext: true, // 丰富的上下文
});
// 自动嵌入存储
await db.storeWithEmbedding({
content: "Your document text",
metadata: { source: "docs", page: 42 }
});
// 查找相似文档
const similar = await db.findSimilar("quantum computing", {
limit: 5,
minScore: 0.75
});
// 结合向量相似性和元数据过滤
const results = await db.hybridSearch({
query: "machine learning models",
filters: {
category: "research",
date: { $gte: "2024-01-01" }
},
limit: 20
});
// 构建 RAG 管道
async function ragQuery(question: string) {
// 1. 获取相关上下文
const context = await db.searchSimilar(
await embed(question),
{ limit: 5, threshold: 0.7 }
);
// 2. 使用上下文生成答案
const prompt = `Context: ${context.map(c => c.text).join('\n')}
Question: ${question}`;
return await llm.generate(prompt);
}
// 高效的批量存储
await db.batchStore(documents.map(doc => ({
text: doc.content,
embedding: doc.vector,
metadata: doc.meta
})));
# 为 Claude Code 启动 AgentDB MCP 服务器
npx agentdb@latest mcp
# 添加到 Claude Code(一次性设置)
claude mcp add agentdb npx agentdb@latest mcp
# 现在可以在 Claude Code 中使用 MCP 工具:
# - agentdb_query: 语义向量搜索
# - agentdb_store: 存储带嵌入的文档
# - agentdb_stats: 数据库统计信息
# 运行全面的基准测试
npx agentdb@latest benchmark
# 结果:
# ✅ 模式搜索:快 150 倍(100µs 对比 15ms)
# ✅ 批量插入:快 500 倍(100 个向量 2ms 对比 1s)
# ✅ 大规模查询:快 12,500 倍(1M 向量 8ms 对比 100s)
# ✅ 内存效率:通过量化减少 4-32 倍
AgentDB 提供多种量化策略以提高内存效率:
const adapter = await createAgentDBAdapter({
quantizationType: 'binary', // 768 维 → 96 字节
});
const adapter = await createAgentDBAdapter({
quantizationType: 'scalar', // 768 维 → 768 字节
});
const adapter = await createAgentDBAdapter({
quantizationType: 'product', // 768 维 → 48-96 字节
});
# 余弦相似度(默认,适用于大多数用例)
npx agentdb@latest query ./db.sqlite "[...]" -m cosine
# 欧几里得距离(L2 范数)
npx agentdb@latest query ./db.sqlite "[...]" -m euclidean
# 点积(用于归一化向量)
npx agentdb@latest query ./db.sqlite "[...]" -m dot
# 检查 HNSW 索引是否启用(自动)
npx agentdb@latest stats ./vectors.db
# 预期:搜索时间 <100µs
# 启用二进制量化(减少 32 倍)
# 在适配器中使用:quantizationType: 'binary'
# 调整相似度阈值
npx agentdb@latest query ./db.sqlite "[...]" -t 0.8 # 更高阈值
# 或使用 MMR 获取多样化结果
# 在适配器中使用:useMMR: true
# 检查嵌入模型维度:
# - OpenAI ada-002: 1536
# - sentence-transformers: 768
# - all-MiniLM-L6-v2: 384
npx agentdb@latest init ./db.sqlite --dimension 768
# 获取全面的统计信息
npx agentdb@latest stats ./vectors.db
# 显示:
# - 总模式/向量数
# - 数据库大小
# - 平均置信度
# - 领域分布
# - 索引状态
npx agentdb@latest mcp 用于 Claude Codenpx agentdb@latest --helpnpx agentdb@latest help <command>每周安装量
0
代码仓库
首次出现
1970年1月1日
安全审计
Implements vector-based semantic search using AgentDB's high-performance vector database with 150x-12,500x faster operations than traditional solutions. Features HNSW indexing, quantization, and sub-millisecond search (<100µs).
# Initialize with default dimensions (1536 for OpenAI ada-002)
npx agentdb@latest init ./vectors.db
# Custom dimensions for different embedding models
npx agentdb@latest init ./vectors.db --dimension 768 # sentence-transformers
npx agentdb@latest init ./vectors.db --dimension 384 # all-MiniLM-L6-v2
# Use preset configurations
npx agentdb@latest init ./vectors.db --preset small # <10K vectors
npx agentdb@latest init ./vectors.db --preset medium # 10K-100K vectors
npx agentdb@latest init ./vectors.db --preset large # >100K vectors
# In-memory database for testing
npx agentdb@latest init ./vectors.db --in-memory
# Basic similarity search
npx agentdb@latest query ./vectors.db "[0.1,0.2,0.3,...]"
# Top-k results
npx agentdb@latest query ./vectors.db "[0.1,0.2,0.3]" -k 10
# With similarity threshold (cosine similarity)
npx agentdb@latest query ./vectors.db "0.1 0.2 0.3" -t 0.75 -m cosine
# Different distance metrics
npx agentdb@latest query ./vectors.db "[...]" -m euclidean # L2 distance
npx agentdb@latest query ./vectors.db "[...]" -m dot # Dot product
# JSON output for automation
npx agentdb@latest query ./vectors.db "[...]" -f json -k 5
# Verbose output with distances
npx agentdb@latest query ./vectors.db "[...]" -v
广告位招租
在这里展示您的产品或服务
触达数万 AI 开发者,精准高效
# Export vectors to JSON
npx agentdb@latest export ./vectors.db ./backup.json
# Import vectors from JSON
npx agentdb@latest import ./backup.json
# Get database statistics
npx agentdb@latest stats ./vectors.db
import { createAgentDBAdapter, computeEmbedding } from 'agentic-flow/reasoningbank';
// Initialize with vector search optimizations
const adapter = await createAgentDBAdapter({
dbPath: '.agentdb/vectors.db',
enableLearning: false, // Vector search only
enableReasoning: true, // Enable semantic matching
quantizationType: 'binary', // 32x memory reduction
cacheSize: 1000, // Fast retrieval
});
// Store document with embedding
const text = "The quantum computer achieved 100 qubits";
const embedding = await computeEmbedding(text);
await adapter.insertPattern({
id: '',
type: 'document',
domain: 'technology',
pattern_data: JSON.stringify({
embedding,
text,
metadata: { category: "quantum", date: "2025-01-15" }
}),
confidence: 1.0,
usage_count: 0,
success_count: 0,
created_at: Date.now(),
last_used: Date.now(),
});
// Semantic search with MMR (Maximal Marginal Relevance)
const queryEmbedding = await computeEmbedding("quantum computing advances");
const results = await adapter.retrieveWithReasoning(queryEmbedding, {
domain: 'technology',
k: 10,
useMMR: true, // Diverse results
synthesizeContext: true, // Rich context
});
// Store with automatic embedding
await db.storeWithEmbedding({
content: "Your document text",
metadata: { source: "docs", page: 42 }
});
// Find similar documents
const similar = await db.findSimilar("quantum computing", {
limit: 5,
minScore: 0.75
});
// Combine vector similarity with metadata filtering
const results = await db.hybridSearch({
query: "machine learning models",
filters: {
category: "research",
date: { $gte: "2024-01-01" }
},
limit: 20
});
// Build RAG pipeline
async function ragQuery(question: string) {
// 1. Get relevant context
const context = await db.searchSimilar(
await embed(question),
{ limit: 5, threshold: 0.7 }
);
// 2. Generate answer with context
const prompt = `Context: ${context.map(c => c.text).join('\n')}
Question: ${question}`;
return await llm.generate(prompt);
}
// Efficient batch storage
await db.batchStore(documents.map(doc => ({
text: doc.content,
embedding: doc.vector,
metadata: doc.meta
})));
# Start AgentDB MCP server for Claude Code
npx agentdb@latest mcp
# Add to Claude Code (one-time setup)
claude mcp add agentdb npx agentdb@latest mcp
# Now use MCP tools in Claude Code:
# - agentdb_query: Semantic vector search
# - agentdb_store: Store documents with embeddings
# - agentdb_stats: Database statistics
# Run comprehensive benchmarks
npx agentdb@latest benchmark
# Results:
# ✅ Pattern Search: 150x faster (100µs vs 15ms)
# ✅ Batch Insert: 500x faster (2ms vs 1s for 100 vectors)
# ✅ Large-scale Query: 12,500x faster (8ms vs 100s at 1M vectors)
# ✅ Memory Efficiency: 4-32x reduction with quantization
AgentDB provides multiple quantization strategies for memory efficiency:
const adapter = await createAgentDBAdapter({
quantizationType: 'binary', // 768-dim → 96 bytes
});
const adapter = await createAgentDBAdapter({
quantizationType: 'scalar', // 768-dim → 768 bytes
});
const adapter = await createAgentDBAdapter({
quantizationType: 'product', // 768-dim → 48-96 bytes
});
# Cosine similarity (default, best for most use cases)
npx agentdb@latest query ./db.sqlite "[...]" -m cosine
# Euclidean distance (L2 norm)
npx agentdb@latest query ./db.sqlite "[...]" -m euclidean
# Dot product (for normalized vectors)
npx agentdb@latest query ./db.sqlite "[...]" -m dot
# Check if HNSW indexing is enabled (automatic)
npx agentdb@latest stats ./vectors.db
# Expected: <100µs search time
# Enable binary quantization (32x reduction)
# Use in adapter: quantizationType: 'binary'
# Adjust similarity threshold
npx agentdb@latest query ./db.sqlite "[...]" -t 0.8 # Higher threshold
# Or use MMR for diverse results
# Use in adapter: useMMR: true
# Check embedding model dimensions:
# - OpenAI ada-002: 1536
# - sentence-transformers: 768
# - all-MiniLM-L6-v2: 384
npx agentdb@latest init ./db.sqlite --dimension 768
# Get comprehensive stats
npx agentdb@latest stats ./vectors.db
# Shows:
# - Total patterns/vectors
# - Database size
# - Average confidence
# - Domains distribution
# - Index status
npx agentdb@latest mcp for Claude Codenpx agentdb@latest --helpnpx agentdb@latest help <command>Weekly Installs
0
Repository
First Seen
Jan 1, 1970
Security Audits
超能力技能使用指南:AI助手技能调用优先级与工作流程详解
39,200 周安装