Unverified50% confidenceCautionExact time
FTS5 依赖字面匹配，当查询词与存储内容用词不同但语义相近时可能漏检，是纯全文方案的固有局限
1
Sources
50%
Confidence
Long-term
Relevance
8/26/2026
First Seen
Sources
Related Entities
Related Claims
Unverified对于事实性、关键词导向的智能体记忆场景，基于关键词的 FTS5 检索完全够用，且响应更快、成本几乎为零68% similarUnverifiedHybrid retrieval combining keyword search with vector search can compensate for the coverage blind spots of pure semantic retrieval.66% similarUnverifiedSplitting documents into chunks by a fixed character count destroys semantic integrity; intelligent semantic chunking or a sliding window strategy can ensure that each document chunk contains complete semantic information.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/804545API
curl https://kongchang.com/api/v1/knowledge/claims/804545MCP
get_claim(id=804545)