Unverified50% confidenceFactExact time
"Dream Deep Memory" divides memories into seven categories: skills, soul, users, projects, topics, facts, lessons and guidelines.
1
Sources
50%
Confidence
Long-term
Relevance
11/16/2026
First Seen
Sources
Open-Source Memory Plugin Dream Recall: Making AI Truly Remember You
bilibili大龙家的小墨11/15/2026
Related Entities
Related Claims
UnverifiedThe experiential memory of self-evolving agents can be divided into three categories: skill memory, factual memory and tool memory.71% similarUnverifiedMemory should be layered: short-term memory stores task state, long-term memory accumulates user preferences, MCP addresses how tools are integrated, and Skills address what process tasks follow.65% similarUnverifiedProblems in development are summarized into eight directions: context, memory, tool calling, termination, hallucination, evaluation, cost, and concurrency.64% similarVerifiedThe Hermes Agent manual proposes a five-layer memory architecture, including the immediate memory layer, short-term working memory, episodic memory, semantic memory, and long-term memory.64% similarUnverified吴恩达梳理的智能体工作流五大核心设计模式为:规划、工具使用、反思、多智能体协作和记忆63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1174418API
curl https://kongchang.com/api/v1/knowledge/claims/1174418MCP
get_claim(id=1174418)