diff --git a/README.md b/README.md index 256c009..eededda 100644 --- a/README.md +++ b/README.md @@ -169,6 +169,79 @@ bash scripts/calculate_llm_score.sh eval_benchmark --- +## EverMind Ecosystem + +EverMind connects memory research, production-ready products, and practical +integrations into one open-source ecosystem. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
Products
EverOSA local-first, Markdown-native long-term memory runtime for agents and users.
RavenA memory-first, self-improving agent harness with proactivity, context control, and skill evolution.
EverMe (CLI)A CLI and agent plugin suite for cross-device, cross-agent personal memory.
Research & Evaluation
SkillCorpusCurated, retrieval-ready agent skill corpora with retrieval and evaluation tooling.
EverAlgoStateless extraction, ranking, parsing, and memory operators that power EverOS.
HyperMemHypergraph-based hierarchical memory for coarse-to-fine long-term conversation retrieval.
MSAMemory Sparse Attention for scalable latent memory and 100M-token contexts.
EverMemBenchEvaluation of factual recall, applied reasoning, and personalized generalization in memory systems.
EvoAgentBenchLongitudinal evaluation of agent self-evolution, transfer efficiency, error avoidance, and skill use.
Integrations
OpenClawOpenClaw plugin for automatic recall, capture, and session-memory lifecycle management.
Hermes AgentHermes plugin for persistent memory across Hermes sessions.
DeepSeek HarnessDSH plugin for memory-aware DeepSeek Harness agents.
DifySelf-hosted and cloud tools for explicit memory search and storage in workflows and agents.
+ +Together, these projects form EverMind's research-to-runtime stack: methods +and benchmarks become reusable memory infrastructure, products, and agent +integrations. + ## Citation ```bibtex diff --git a/README_ZH.md b/README_ZH.md index 35a1de1..1a64dbb 100644 --- a/README_ZH.md +++ b/README_ZH.md @@ -168,6 +168,77 @@ bash scripts/calculate_llm_score.sh eval_benchmark --- +## EverMind 生态 + +EverMind 将记忆研究、可直接使用的产品与实际集成连接为一个开源生态。 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
产品
EverOS本地优先、Markdown 原生的 Agent 与用户长期记忆运行时。
Raven以记忆为核心的自进化 Agent Harness，具备主动性、上下文控制与 Skill 进化能力。
EverMe（CLI）面向跨设备、跨 Agent 个人记忆的 CLI 与 Agent 插件套件。
研究与评测
SkillCorpus将分散的 Agent Skill 整理为可检索语料库，并提供检索与评测工具。
EverAlgo为 EverOS 提供无状态的提取、排序、解析与记忆算法。
HyperMem基于超图的分层记忆架构，用于由粗到细的长期对话检索。
MSA面向可扩展潜在记忆与一亿 Token 上下文的 Memory Sparse Attention。
EverMemBench从事实召回、应用推理和个性化泛化三个层面评测记忆系统。
EvoAgentBench纵向评测 Agent 自进化、迁移效率、错误规避和 Skill 使用能力。
插件与集成
OpenClawOpenClaw 插件，自动管理召回、写入与会话记忆生命周期。
Hermes AgentHermes 插件，为 Hermes 会话提供持久记忆。
DeepSeek HarnessDSH 插件，让 DeepSeek Harness Agent 使用长期记忆。
Dify本地版与云端版工具，在工作流和 Agent 中显式搜索与写入记忆。
+ +这些项目共同构成 EverMind 从研究到运行时的完整链路：将方法与评测转化为 +可复用的记忆基础设施、产品和 Agent 集成。 + ## 引用 ```bibtex