Skip to content

docs: explain upcoming knowledge wiki and reflection #289

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
cyfyifanchen merged 10 commits into from
Jun 16, 2026
+185 −48
Merged

docs: explain upcoming knowledge wiki and reflection #289

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
06fc195
docs: explain upcoming knowledge wiki and reflection
cyfyifanchen Jun 16, 2026
0b7fd0e
docs: tighten roadmap summary in readmes
cyfyifanchen Jun 16, 2026
3ceb528
docs: add memory library comparison table
cyfyifanchen Jun 16, 2026
020aa3d
docs: make memory comparison a named matrix
cyfyifanchen Jun 16, 2026
215abe2
docs: use category comparison for memory positioning
cyfyifanchen Jun 16, 2026
c1e13ca
docs: sharpen readme comparison table
cyfyifanchen Jun 16, 2026
2eaac5a
docs: refine everme readme positioning
cyfyifanchen Jun 16, 2026
639b44f
docs: correct readme positioning title to everos
cyfyifanchen Jun 16, 2026
629b69a
docs: simplify readme comparison table
cyfyifanchen Jun 16, 2026
c4d5e3e
docs: clarify reflection idle behavior
cyfyifanchen Jun 16, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
120 changes: 95 additions & 25 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,8 +21,9 @@

<br>

- [EverOS 1.0.0 Highlights](#everos-100-highlights)
- [Why EverOS](#why-everos)
- [EverOS 1.0.0](#everos-100)
- [EverOS: One Memory For All](#everos-one-memory-for-all)
- [How EverOS Is Different](#how-everos-is-different)
- [Quick Start](#quick-start)
- [Architecture At A Glance](#architecture-at-a-glance)
- [Storage Layout](#storage-layout)
Expand All @@ -39,7 +40,7 @@
</details>


## EverOS 1.0.0 Highlights
## EverOS 1.0.0

> [!IMPORTANT]
>
Expand All @@ -48,8 +49,26 @@
> multimodal ingestion, user and agent memory scopes, and modular algorithms
> through [EverAlgo](https://github.com/EverMind-AI/EverAlgo).
>
> **Watch this repository** for the next wave of memory-system work, including
> Wiki-style knowledge layers and Dreaming for deeper offline evolution.
> **Coming next:** Knowledge Wiki will turn memory into editable,
> source-backed Markdown knowledge pages. Reflection, also called Dreaming,
> will run when the system is idle or offline to connect signals, compress
> history, and improve profiles and skills between sessions.

<br>
<div align="right">

[![](https://img.shields.io/badge/-Back_to_top-gray?style=flat-square)](#readme-top)

</div>


## EverOS: One Memory For All

EverOS is the local memory operating system for agents and makers. It gives
one portable memory layer across coding assistants, apps, devices, and
workflows. Today it stores conversations, files, and agent trajectories as
readable Markdown, then syncs local SQLite and LanceDB indexes for fast
retrieval and self-evolving reuse.

<table>
<tr>
Expand Down Expand Up @@ -102,24 +121,50 @@ Search independently by <code>user_id</code>, <code>agent_id</code>,
</div>


## Why EverOS

EverOS is an open-source Python framework for self-evolving long-term
memory across agents and platforms. It gives makers one portable memory
layer for every agent they use - Claude Code, Codex, OpenClaw, Hermes,
and more - so context, decisions, files, and trajectories can follow the
work instead of staying trapped in one tool.
## How EverOS Is Different

EverOS stores conversations, agent trajectories, and files as readable
Markdown, then syncs local SQLite and LanceDB indexes for fast retrieval.
Agents can reuse past cases and skills, improve from repeated workflows,
and become more proactive over time.

The system is built around three boundaries:

1. **Memory content stays readable** - Markdown is the durable source of truth.
2. **Runtime state stays local** - SQLite tracks state and LanceDB handles vector, BM25, and scalar-filter search.
3. **Algorithms stay modular** - [EverAlgo](https://github.com/EverMind-AI/EverAlgo) owns memory algorithms; EverOS owns runtime, persistence, online flows, and offline evolution.
<table>
<tr>
<th width="28%">Title</th>
<th width="36%">EverOS</th>
<th width="36%">Other Agent Memory Libraries</th>
</tr>
<tr>
<td><strong>Markdown source of truth</strong></td>
<td>✅ Canonical <code>.md</code> files that are readable, editable, diffable, and Git-versioned</td>
<td>❌ Usually API, vector, graph, dashboard, or database state</td>
</tr>
<tr>
<td><strong>Direct file editing</strong></td>
<td>✅ Edit <code>.md</code> files; cascade watcher syncs</td>
<td>❌ Usually SDK, API, dashboard, or backend update paths</td>
</tr>
<tr>
<td><strong>Local three-part stack</strong></td>
<td>✅ Markdown + SQLite + LanceDB; no MongoDB, Elasticsearch, or Redis required</td>
<td>❌ Often depends on managed services, vector DBs, graph DBs, or server stacks</td>
</tr>
<tr>
<td><strong>User + agent tracks</strong></td>
<td>✅ User <code>episodes/profile</code> and agent <code>cases/skills</code> are separate first-class surfaces</td>
<td>❌ Usually centered on chat history, profiles, entities, facts, or retrieval records</td>
</tr>
<tr>
<td><strong>Orthogonal retrieval</strong></td>
<td>✅ Search by <code>user_id</code>, <code>agent_id</code>, <code>app_id</code>, <code>project_id</code>, and <code>session_id</code></td>
<td>❌ Usually app, namespace, tenant, thread, or graph scoped</td>
</tr>
<tr>
<td><strong>Knowledge Wiki</strong></td>
<td>✅ Coming next: editable, source-backed Markdown knowledge pages built from memory</td>
<td>❌ Usually retrieval, graph, dashboards, or generated summaries instead of editable source-backed pages</td>
</tr>
<tr>
<td><strong>Dreaming / Reflection</strong></td>
<td>✅ Coming next: Reflection that runs when the system is idle or offline to connect signals, compress history, and improve profiles and skills between sessions</td>
<td>❌ Usually online read/write APIs, retrieval records, or summaries rather than idle-time memory consolidation</td>
</tr>
</table>

<br>
<div align="right">
Expand Down Expand Up @@ -665,9 +710,34 @@ Explore stored entities and relationships in a graph interface. Frontend demo; b
## Watch EverOS

EverOS 1.0.0 is the first release of a larger memory-system roadmap.
Watch this repository for upcoming work on Wiki-style memory, Dreaming,
deeper offline evolution, benchmark releases, and more real-world agent
integrations.
Watch this repository for upcoming work on deeper idle-time and offline evolution,
benchmark releases, and more real-world agent integrations.

<table>
<tr>
<td width="50%" valign="top">
<strong>Knowledge Wiki</strong><br>
<br>
Turns scattered episodes, files, facts, and agent traces into source-backed
Markdown pages for people, projects, topics, decisions, and workflows. Memory
becomes something users can read, correct, link, version, and open in their
existing Markdown tools.
</td>
<td width="50%" valign="top">
<strong>Dreaming / Reflection</strong><br>
<br>
Runs when the system is idle or offline to revisit stored memory, connect weak
signals, compress noisy history into durable patterns, and improve profiles and
skills. The agent gets better between active sessions, not only while you prompt
it.
</td>
</tr>
</table>

Most memory systems stop at chat history, opaque profiles, or vector recall.
EverOS keeps memory local, Markdown-native, auditable, and self-evolving: raw
memory stays readable, derived knowledge becomes a wiki, and Reflection turns
repeated experience into more useful long-term behavior.

If EverOS is useful to your agent stack, starring the repo helps more
builders discover it.
Expand Down
113 changes: 90 additions & 23 deletions README.zh-CN.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
<div align="center" id="readme-top">

![EverOS banner](https://github.com/EverMind-AI/EverOS/releases/download/v1.0.0/everos-readme-banner-optimized.jpg)
![EverOS banner](https://github.com/user-attachments/assets/8e217d39-5d15-4c6c-9b54-3e83add4e0f2)

<p align="center">
<a href="https://x.com/evermind"><img src="https://img.shields.io/badge/EverMind-000000?labelColor=gray&style=for-the-badge&logo=x&logoColor=white" alt="X"></a>
Expand All @@ -21,8 +21,9 @@

<br>

- [EverOS 1.0.0 亮点](#everos-100-亮点)
- [为什么选择 EverOS](#为什么选择-everos)
- [EverOS 1.0.0](#everos-100)
- [EverOS: One Memory For All](#everos-one-memory-for-all)
- [EverOS 的差异](#everos-的差异)
- [快速开始](#快速开始)
- [架构概览](#架构概览)
- [存储布局](#存储布局)
Expand All @@ -39,7 +40,7 @@
</details>


## EverOS 1.0.0 亮点
## EverOS 1.0.0

> [!IMPORTANT]
>
Expand All @@ -48,8 +49,24 @@
> 多模态摄取、用户记忆与 Agent 记忆作用域，以及由
> [EverAlgo](https://github.com/EverMind-AI/EverAlgo) 支撑的模块化算法。
>
> **欢迎 Watch 这个仓库。** 下一阶段我们会继续推进记忆系统方法，
> 包括 Wiki 式知识层和用于更深层离线进化的 Dreaming。
> **即将推出：** Knowledge Wiki 会把记忆整理成可编辑、可溯源的
> Markdown 知识页。Reflection（也称 Dreaming）会在系统空闲或离线时
> 连接信号、压缩历史，并持续改进 profile 和 skills。

<br>
<div align="right">

[![](https://img.shields.io/badge/-Back_to_top-gray?style=flat-square)](#readme-top)

</div>


## EverOS: One Memory For All

EverOS 是面向 agents 和 makers 的本地记忆操作系统。它提供一层可携带的
统一记忆层，让记忆穿过 coding assistants、apps、devices 和 workflows。
目前它会把对话、文件和 Agent 轨迹保存为可读 Markdown，并同步本地 SQLite
和 LanceDB 索引，用于快速检索和自进化复用。

<table>
<tr>
Expand Down Expand Up @@ -96,22 +113,50 @@ Agent 记忆（<code>cases</code> / <code>skills</code>）与用户记忆（<cod
</div>


## 为什么选择 EverOS

EverOS 是一个开源 Python 框架，用来构建**跨 Agent、跨平台的自进化长期记忆**。
它为 maker 提供一层可携带的统一记忆层，适用于他们使用的每一个 Agent：
Claude Code、Codex、OpenClaw、Hermes 等等。这样，上下文、决策、文件和
Agent 轨迹可以跟着工作流走，而不是被锁在某一个工具里。
## EverOS 的差异

EverOS 会把对话、Agent 轨迹和文件保存为可读 Markdown，并同步本地 SQLite
和 LanceDB 索引，以便快速检索。Agent 可以复用过去的 cases 和 skills，从重复
工作流中自我改进，并逐渐变得更加主动。

系统围绕三个边界设计：

1. **记忆内容保持可读** - Markdown 是长期、耐用的 source of truth。
2. **运行时状态保持本地** - SQLite 跟踪状态；LanceDB 处理向量、BM25 和结构化过滤搜索。
3. **算法保持模块化** - [EverAlgo](https://github.com/EverMind-AI/EverAlgo) 负责记忆算法；EverOS 负责运行时、持久化、在线流程和离线进化。
<table>
<tr>
<th width="28%">Title</th>
<th width="36%">EverOS</th>
<th width="36%">Other Agent Memory Libraries</th>
</tr>
<tr>
<td><strong>Markdown source of truth</strong></td>
<td>✅ 标准 <code>.md</code> 文件：可读、可编辑、可 diff、可 Git 版本化</td>
<td>❌ 通常是 API、vector、graph、dashboard 或 database state</td>
</tr>
<tr>
<td><strong>直接文件编辑</strong></td>
<td>✅ 编辑 <code>.md</code>；cascade watcher 同步</td>
<td>❌ 通常需要 SDK、API、dashboard 或 backend update path</td>
</tr>
<tr>
<td><strong>本地三件套</strong></td>
<td>✅ Markdown + SQLite + LanceDB；不需要 MongoDB、Elasticsearch 或 Redis</td>
<td>❌ 常依赖 managed service、vector DB、graph DB 或 server stack</td>
</tr>
<tr>
<td><strong>用户 + Agent 双轨</strong></td>
<td>✅ 用户 <code>episodes/profile</code> 与 Agent <code>cases/skills</code> 是分离的一等记忆表面</td>
<td>❌ 通常围绕 chat history、profiles、entities、facts 或 retrieval records</td>
</tr>
<tr>
<td><strong>正交检索作用域</strong></td>
<td>✅ 按 <code>user_id</code>、<code>agent_id</code>、<code>app_id</code>、<code>project_id</code> 和 <code>session_id</code> 检索</td>
<td>❌ 通常按 app、namespace、tenant、thread 或 graph 来组织</td>
</tr>
<tr>
<td><strong>Knowledge Wiki</strong></td>
<td>✅ 即将推出：由记忆形成可编辑、可溯源的 Markdown 知识页</td>
<td>❌ 通常是 retrieval、graph、dashboard 或 generated summaries，而不是可编辑、可溯源的知识页</td>
</tr>
<tr>
<td><strong>Dreaming / Reflection</strong></td>
<td>✅ 即将推出：在系统空闲或离线时运行，用来连接信号、压缩历史，并在 session 之间改进 profiles 和 skills</td>
<td>❌ 通常是在线读写 API、retrieval records 或 summaries，而不是空闲态记忆整理</td>
</tr>
</table>

<br>
<div align="right">
Expand Down Expand Up @@ -651,8 +696,30 @@ Claude Code 的持久记忆插件。自动保存并回忆过去 coding sessions
## 关注 EverOS

EverOS 1.0.0 是更大规模记忆系统路线图的第一个发布版本。Watch 这个仓库，
即可持续关注 Wiki 式记忆、Dreaming、更深入的离线进化、benchmark releases，
以及更多真实 Agent 集成。
即可持续关注更深入的空闲态和离线进化、benchmark releases，以及更多真实 Agent 集成。

<table>
<tr>
<td width="50%" valign="top">
<strong>Knowledge Wiki</strong><br>
<br>
把分散的 episodes、files、facts 和 Agent traces 整理成有来源的 Markdown
知识页，覆盖 people、projects、topics、decisions 和 workflows。记忆不再只是
向量召回结果，而是用户可以阅读、修正、链接、版本化，并用现有 Markdown 工具打开的知识层。
</td>
<td width="50%" valign="top">
<strong>Dreaming / Reflection</strong><br>
<br>
在系统空闲或离线时运行，重新审视已存储记忆，连接弱信号，把嘈杂历史压缩成稳定模式，
并持续改进 profile 和 skills。目标是让 Agent 在活跃 session 之间也能变得更好，
而不是只在你 prompt 它时才进步。
</td>
</tr>
</table>

许多记忆系统停留在聊天历史、黑盒 profile 或向量召回。EverOS 的差异在于：
记忆保持本地、Markdown-native、可审计、可自进化；原始记忆仍然可读，
衍生知识沉淀为 wiki，Reflection 则把重复经验转化为更有用的长期行为。

如果 EverOS 对你的 Agent stack 有帮助，Star 这个仓库也会帮助更多 builders
发现它。
Expand Down
Loading