AETHMERE-OS · LOCAL CONTEXT FOR YOUR AI

Aethmere-OS让你的 AI，接得上工作。

用桌面端整理你希望 AI 记住的项目背景，也可以通过 MCP 接给 Codex 或 Claude Code。少重复解释，重要内容按需读取；项目文件不上传，云端治理只在已配置环境中接收最小会话文本。

接入 Codex / Claude Code 下载 Windows 桌面端 ↓
Agent Client：Windows / macOS / Linux · Agent Studio：Windows x64
本地项目上下文03
目标约束决定
MCP 已连接

先读取相关背景，再继续当前任务。

PROJECT_GOALKEY_DECISIONOPEN_QUESTION
项目文件仍在本机Codex · Claude Code
WHAT YOU GET

不是多一个聊天框，而是给现有 AI 补上项目上下文。

你继续使用熟悉的 AI 客户端。Aethmere 只负责把你明确保存的背景，以可见、可控的方式接进去。

01

背景不用反复讲

把目标、约束和关键决定留在项目里。下一次打开 AI，它仍能接着这件事工作。

连续上下文
02

重要内容能精确找到

每条上下文都有稳定 ID。智能体可以先列出、再按 ID 读取，而不是把整个项目一次性塞进对话。

按需读取
03

项目文件保留在本机

Agent Client 不联网；Studio 的会话与偏好保存在本机。只有已配置云端治理时，Studio 才发送完成治理所需的最小会话文本，不发送项目路径、文件、Git、Shell 或令牌。

边界可核验
START HERE · 5 MINUTES

先确定用途，再安装对应组件。

日常接入 Codex 或 Claude Code，请安装 Agent Client；想在 Windows 上与本机 Ollama 模型对话，请使用 Agent Studio。评测 CLI 只用于复核公开数据，不参与日常接入。

推荐 · 接入现有 AIAGENT CLIENT · NODE 20+

把当前项目接给 Codex 或 Claude Code

这是大多数用户需要的安装路径。打开 PowerShell、Windows Terminal 或 macOS / Linux 终端，按顺序完成四步。

  1. 01

    确认 Node.js 已就绪

    node --version
npm --version
    需要 Node.js 20 或更高版本；新安装建议使用当前 LTS 版本。 前往 Node.js 官网 ↗
  2. 02

    从官网安装 Agent Client

    npm install --global https://aethmere.com/downloads/aethmere-agent-0.10.0.tgz
aethmere-agent --version
    看到 “Aethmere Agent Client 0.10.0” 就表示安装成功。
  3. 03

    在项目里保存第一条上下文

    cd "你的项目目录"
aethmere-agent init
aethmere-agent add --id PROJECT_GOAL --title "项目目标" --text "让 AI 接得上这个项目的目标和约束"
aethmere-agent list
    把第一行替换成真实项目路径。内容只写入该项目的 .aethmere/context.json.
  4. 04

    先预览，再写入 MCP 配置

    aethmere-agent doctor
aethmere-agent connect --client all --check
aethmere-agent connect --client all
    只使用一个客户端时，可把 all 换成 codex 或 claude。最后一条命令执行后，重新打开对应客户端；已有配置会先保留备份。
独立桌面聊天 · WindowsWINDOWS X64

使用 Agent Studio 桌面端

Agent Studio 不是 Codex / Claude Code 的接入组件。它需要本机先安装并运行 Ollama，且 Ollama 中至少已有一个模型；未配置 Aethmere 云端治理时，仍可进行本地对话。 下载 Ollama for Windows ↗

  1. 01
    先准备 Ollama 和本地模型确认 Ollama 正在运行，并已下载至少一个可用模型。
  2. 02
    下载并完整解压 ZIP不要直接在压缩包里运行。
  3. 03
    打开 Aethmere Agent Studio.exe当前是未签名的便携预览版，Windows 可能显示“未知发布者”。
  4. 04
    选择本地模型，项目可选会话、模型和项目选择保存在本机；不会自动扫描项目。
下载 Windows 桌面端
可选工具 · 不参与日常接入

只在需要复核公开评测时安装

aethmere 是独立的评测工具，用于查看历史 V3/V5 汇总、运行公开证据 ID 示例，并核对官网与 GitHub 的发行文件。如果你只是使用 Aethmere，可以跳过这一段。

npm install --global https://aethmere.com/downloads/aethmere-cli-0.10.2.tgz
aethmere --version
aethmere doctor --online
aethmere eval
aethmere trial
遇到问题？展开查看常见处理方法

提示找不到 node 或 npm： 安装 Node.js 20 或更高版本，然后关闭并重新打开终端。

提示找不到 aethmere-agent： 先重新打开终端；Windows PowerShell 若阻止脚本，可运行 aethmere-agent.cmd --version，或改用“命令提示符”。

connect 显示 not-installed： 表示没有检测到对应客户端的配置。请先安装并打开一次该客户端，或把 --client all 改为你实际使用的 codex 或 claude。

担心命令会上传项目： Agent Client 没有网络请求或遥测，上下文保存在当前项目的 .aethmere/context.json。

想卸载： 运行 npm uninstall --global aethmere-agent aethmere-cli。卸载不会删除项目里的上下文文件。

SEALED V5 · SAME QUESTIONS

同一 7B 模型、同一套题：启用 Aethmere 前后相差多少？

两组使用同一 7B 模型、同一批封存题、相同语言和分母；区别是前者启用 Aethmere，后者不使用记忆并直接回答。这是 2026 年 8 月的历史封存评测，固定题目按确定性答案逐题计分；失败、超时、缺失，以及规定必须回答却没有作答的情况都留在分母中。

同一 7B · 启用 AETHMERE · V5同一 7B 模型
4,800 / 4,800

100.0% 正确 · 0 缺失

+58.7个百分点
同一 7B · 无记忆同一模型 · 直接回答
1,982 / 4,800

41.3% 正确 · 2,818 失败

V5 中英文逐语言结果
题集同一 7B · 启用 Aethmere同一 7B · 无记忆通过率差
中文 · 2,400 题2,400 / 2,400100.0%1,346 / 2,40056.1%+43.9个百分点
English · 2,400 questions2,400 / 2,400100.0%636 / 2,40026.5%+73.5个百分点
合计 · 4,800 题4,800 / 4,800100.0%1,982 / 4,80041.3%+58.7个百分点
更早的 V3 封存集2,400 / 2,4000 缺失
同一 7B 无记忆失败项2,818 / 2,818启用 Aethmere 后在 V5 中答对
同一 7B 无记忆原本正确的题0 回退启用 Aethmere · V5 中英文合计
下载公开汇总 ↓方法与限制 ↗

这些是受限任务的历史封存结果，不代表当前生产版本、这个公开客户端、开放世界问答或任意问题的准确率。公开文件只有汇总与完整性哈希，不含私有题目、提示词和原始模型输出。

18-DIMENSION NATIVE SCORECARD · V13 · 2026-08-26

V5 实测综合分已重算：Aethmere 10.00 / 10，18 项全部稳定通过。

五个锁定版本在同一 18 维合同下各跑两轮，共核对 180 格。综合分只按两轮稳定 PASS 计：10 × 稳定 PASS ÷ 18；FAIL、FLAKY 与 N/A 都不加分。旧版主观文档底分与权重已撤下。

Aethmere0.7.0 · 2c1df71
10.00 / 10第 1 / 5 · 18 / 18 稳定 PASS
Letta Code0.30.29
7.22 / 10第 2 / 5 · 13 / 18 稳定 PASS
MemOS2.0.30
7.22 / 10第 2 / 5 · 13 / 18 稳定 PASS
Mem02.0.18
6.67 / 10第 4 / 5 · 12 / 18 稳定 PASS
Graphiti0.29.3
0.56 / 10第 5 / 5 · 1 / 18 稳定 PASS
18 维两轮稳定结果 · PASS 必须两轮都通过且语义一致
维度AethmereLetta CodeMemOSMem0Graphiti
D01原生写入、持久标识、精确读取PASSPASSPASSPASSPASS
D02定址检索与作用域隔离PASSPASSPASSPASSFAIL
D03排序相关性检索PASSFAILFAILPASSFAIL
D04更新与当前值PASSPASSPASSPASSFAIL
D05生命周期转换与移除PASSPASSPASSPASSFAIL
D06逻辑作用域隔离PASSPASSFAILPASSFAIL
D07数据根隔离PASSPASSPASSFAILFAIL
D08元数据过滤PASSN/AFAILPASSFAIL
D09时序边界PASSPASSPASSPASSFAIL
D10版本历史PASSPASSFAILPASSFAIL
D11关系解析PASSFAILFAILFAILFAIL
D12来源追踪PASSPASSPASSPASSFAIL
D13批量持久化与恰好一次PASSFAILPASSPASSFAIL
D14重启恢复PASSPASSPASSFAILFLAKY
D15官方导出PASSPASSPASSN/AN/A
D16官方导入PASSPASSPASSN/AN/A
D17并发、幂等与完整性PASSFAILPASSFLAKYFAIL
D18非法输入隔离与恢复PASSPASSPASSPASSFAIL
稳定 PASS / 适用维度（N/A 不计入适用维度分母）
产品锁定版本稳定 PASSFAILFLAKYN/A
Aethmere0.7.0 · 2c1df7118 / 18000
Letta Code0.30.2913 / 17401
MemOS2.0.3013 / 18500
Mem02.0.1812 / 16312
Graphiti0.29.31 / 161412

10.00 只表示 Aethmere 在这份锁定 V5 合同中 18 / 18 两轮稳定通过，不是任意产品质量满分。Graphiti 的 14 个稳定 FAIL 是原生操作未在锁定时限内完成，不能据此断言功能不存在。本评测由另一位操作者在独立物理 Apple Silicon Mac 上执行，属于 owner-run，不是第三方认证、开放世界总榜或生产 SLA。

下载 V13 逐维结果与公式 ↓方法、哈希与限制 ↗
PUBLIC DOWNLOADS

四个下载用途不同，不需要全部安装。

接入 Codex 或 Claude Code，只需 Agent Client；独立桌面聊天使用 Agent Studio；VS Code 插件是可选辅助；评测 CLI 只用于复核公开数据。所有文件均从官网直接下载并公开 SHA-256。

推荐NODE · TGZ

Agent Client

在本机保存项目上下文，并通过 MCP 接给 Codex 或 Claude Code。日常接入只需安装这个组件；它没有网络请求或遥测。

npm install -g https://aethmere.com/downloads/aethmere-agent-0.10.0.tgz
下载 TGZ 包 ↓GitHub Release ↗
aethmere-agent-0.10.0.tgz · 6.9 KB
独立桌面聊天WINDOWS X64 · ZIP

Agent Studio

Windows 便携桌面端，需要本机已安装并运行 Ollama。它不替代 Codex / Claude Code 的 Agent Client；不会自动扫描项目，也没有遥测。

完整解压 ZIP 后运行 Aethmere Agent Studio.exe
下载 Windows 版 ↓核对 SHA-256 ↗
aethmere-agent-studio-0.11.0-win32-x64.zip · 143.2 MB · 未签名预览版
可选VS CODE · VSIX

VS Code 插件

可选辅助工具：把编辑器中选中的背景保存到本地上下文、查看条目，并复制当前项目的 MCP 接入命令。下载后在 VS Code 的扩展面板中选择“从 VSIX 安装”。

扩展 → … → 从 VSIX 安装
下载插件 ↓GitHub Release ↗
aethmere-vscode-0.10.0.vsix · 13.0 KB
验证数据NODE · TGZ

评测 CLI

仅用于复核历史 V3/V5 汇总、运行公开证据 ID 示例，并核对官网与 GitHub 的发行哈希；不用于接入 Codex 或 Claude Code。

npm install -g https://aethmere.com/downloads/aethmere-cli-0.10.2.tgz
下载评测 CLI ↓GitHub Release ↗
aethmere-cli-0.10.2.tgz · 5.9 KB
GOOD FIT

适合那些不能只聊一次的事。

如果一次对话就能解决，不需要多一层工具；真正长期、反复推进的工作，才值得把背景沉淀下来。

01
长期研究

保存研究目标、判断依据和仍待确认的问题。

02
复杂项目

让需求、限制和关键决定在多次会话之间接得上。

03
关键判断

用稳定 ID 指向实际可见的上下文，减少凭空补全。

YOUR CONTEXT, YOUR BOUNDARY

哪些内容留在本机，哪些情况下会联网？

不同组件的网络边界不同：Agent Client 完全本地；Agent Studio 默认连接本机 Ollama，只有已配置云端治理时才会发送最小会话文本。

01Agent Client 不联网、不含遥测；项目上下文保存在当前项目的 .aethmere/context.json
02Studio 的会话、模型和项目选择保存在本机；已配置治理时只发送最小会话文本，不发送项目路径、文件、Git、Shell 或令牌
03每个发行文件都有版本、大小和 SHA-256
ABOUT AETHMERE-OS

Aethmere-OS，由青岛果冻先生网络科技有限责任公司开发与运营。

我们围绕长期 AI 工作中的项目上下文、可核验记忆与本地数据边界持续研发 Aethmere-OS（识宙）。本官网提供产品说明、公开下载、评测方法与安全反馈入口。

START LOCALLY

从一个真实项目开始，让 Codex 或 Claude Code 先读到你保存的背景。

查看 5 分钟接入步骤
AETHMERE-OS · LOCAL CONTEXT FOR YOUR AI

Aethmere-OSHelp your AIpick up where work left off.

Organize the project background you want AI to retain in the desktop app, or connect it to Codex and Claude Code through MCP. Explain less, retrieve important context only when needed, and keep project files local. Cloud governance receives only the minimum conversation text in configured environments.

Connect Codex / Claude Code Download the Windows desktop app ↓
Agent Client: Windows / macOS / Linux · Agent Studio: Windows x64
Local project context03
GoalConstraintsDecisions
MCP connected

Read the relevant context, then continue the current task.

PROJECT_GOALKEY_DECISIONOPEN_QUESTION
Project files stay localCodex · Claude Code
WHAT YOU GET

Not another chat box—project context for the AI tools you already use.

Keep using your preferred AI client. Aethmere connects only the background you explicitly save, in a visible and controllable way.

01

Stop repeating the background

Keep goals, constraints, and key decisions in the project. The next time you open your AI tool, it can continue the same work.

Continuous context
02

Find important context precisely

Every context item has a stable ID. Agents can list entries first and then read by ID instead of pushing the whole project into one conversation.

Read on demand
03

Keep project files on your machine

Agent Client is offline. Studio stores conversations and preferences locally; only configured cloud-governance environments send the minimum conversation text needed for governance—never project paths, files, Git data, shell data, or tokens.

Verifiable boundary
START HERE · 5 MINUTES

Choose the job first, then install the matching component.

For everyday use with Codex or Claude Code, install Agent Client. For a standalone Windows chat app backed by local Ollama models, use Agent Studio. The evaluation CLI only verifies public data and is not part of everyday setup.

Recommended · Existing AI toolsAGENT CLIENT · NODE 20+

Connect the current project to Codex or Claude Code

This is the setup path most users need. Open PowerShell, Windows Terminal, or a macOS/Linux terminal and complete these four steps in order.

  1. 01

    Check that Node.js is ready

    node --version
npm --version
    Node.js 20 or later is required. For a new installation, use the current LTS release. Open the Node.js website ↗
  2. 02

    Install Agent Client from the official site

    npm install --global https://aethmere.com/downloads/aethmere-agent-0.10.0.tgz
aethmere-agent --version
    Seeing “Aethmere Agent Client 0.10.0” confirms a successful installation.
  3. 03

    Save the first context item in your project

    cd "path/to/your-project"
aethmere-agent init
aethmere-agent add --id PROJECT_GOAL --title "Project goal" --text "Give the AI the goals and constraints needed to continue this project"
aethmere-agent list
    Replace the first line with the real project path. Content is written only to this project's .aethmere/context.json.
  4. 04

    Preview first, then write the MCP configuration

    aethmere-agent doctor
aethmere-agent connect --client all --check
aethmere-agent connect --client all
    If you use only one client, replace all with codex or claude. After the final command, restart that client; existing configuration is backed up first.
Standalone desktop chat · WindowsWINDOWS X64

Use the Agent Studio desktop app

Agent Studio is not the Codex / Claude Code integration component. Ollama must already be installed and running locally with at least one model available. Local chat still works when Aethmere cloud governance is not configured. Download Ollama for Windows ↗

  1. 01
    Prepare Ollama and a local modelConfirm that Ollama is running and that at least one model has already been downloaded.
  2. 02
    Download and fully extract the ZIPDo not run it from inside the archive.
  3. 03
    Open Aethmere Agent Studio.exeThis is currently an unsigned portable preview, so Windows may show an “Unknown publisher” warning.
  4. 04
    Choose a local model; a project is optionalConversations, model selection, and project selection remain local. The app does not scan projects automatically.
Download the Windows desktop app
Optional tool · Not part of everyday setup

Install this only to verify the public evaluation

aethmere is a separate verification tool. It displays the historical V3/V5 aggregate, runs public evidence-ID examples, and verifies release files on the website and GitHub. Skip this section if you only want to use Aethmere.

npm install --global https://aethmere.com/downloads/aethmere-cli-0.10.2.tgz
aethmere --version
aethmere doctor --online
aethmere eval
aethmere trial
Having trouble? Open the common fixes

node or npm is not found: Install Node.js 20 or later, then close and reopen the terminal.

aethmere-agent is not found: Reopen the terminal first. If Windows PowerShell blocks scripts, run aethmere-agent.cmd --version or use Command Prompt.

connect reports not-installed: The matching client configuration was not detected. Install and open that client once, or replace --client all with the client you actually use: codex or claude.

Concerned that the command uploads your project? Agent Client makes no network requests and has no telemetry. Context stays in .aethmere/context.json inside the current project.

Want to uninstall? Run npm uninstall --global aethmere-agent aethmere-cli. Uninstalling does not remove context files from your projects.

SEALED V5 · SAME QUESTIONS

Same 7B model, same questions: what changed when Aethmere was enabled?

Both groups use the same 7B model, sealed questions, languages, and denominator. One group has Aethmere enabled; the baseline uses no memory and answers directly. This is a historical sealed evaluation from August 2026. Deterministic answers score each fixed item, while failures, timeouts, missing outputs, and required-but-absent answers remain in the denominator.

SAME 7B · AETHMERE ENABLED · V5Same 7B model
4,800 / 4,800

100.0% correct · 0 missing

+58.7percentage points
SAME 7B · NO MEMORYSame model · Direct answers
1,982 / 4,800

41.3% correct · 2,818 failed

V5 results by language
Test setSame 7B · Aethmere enabledSame 7B · No memoryPass-rate gap
Chinese · 2,400 questions2,400 / 2,400100.0%1,346 / 2,40056.1%+43.9percentage points
English · 2,400 questions2,400 / 2,400100.0%636 / 2,40026.5%+73.5percentage points
Combined · 4,800 questions4,800 / 4,800100.0%1,982 / 4,80041.3%+58.7percentage points
Earlier V3 sealed set2,400 / 2,4000 missing
No-memory 7B failures2,818 / 2,818answered correctly with Aethmere enabled in V5
Items the no-memory 7B originally answered correctly0 regressionsAethmere enabled · Combined Chinese and English V5
Download public summary ↓Method and limitations ↗

These are historical sealed results for bounded tasks. They do not represent the current production version, this public client, open-world question answering, or accuracy on arbitrary questions. Public files contain only aggregate results and integrity hashes—not private items, prompts, or raw model outputs.

18-DIMENSION NATIVE SCORECARD · V13 · 2026-08-26

V5 measured score recalculated: Aethmere is 10.00 / 10 with 18 stable passes.

Five locked versions each ran twice under one 18-dimension contract, for 180 checked cells. The composite uses only stable two-run PASS evidence: 10 × stable PASS ÷ 18. FAIL, FLAKY, and N/A earn no points. The former subjective documentation baselines and weights have been withdrawn.

Aethmere0.7.0 · 2c1df71
10.00 / 10Rank 1 / 5 · 18 / 18 stable PASS
Letta Code0.30.29
7.22 / 10Rank 2 / 5 · 13 / 18 stable PASS
MemOS2.0.30
7.22 / 10Rank 2 / 5 · 13 / 18 stable PASS
Mem02.0.18
6.67 / 10Rank 4 / 5 · 12 / 18 stable PASS
Graphiti0.29.3
0.56 / 10Rank 5 / 5 · 1 / 18 stable PASS
Stable results across 18 dimensions · PASS requires both runs to pass with matching semantics
DimensionAethmereLetta CodeMemOSMem0Graphiti
D01Native write, durable identity, exact readPASSPASSPASSPASSPASS
D02Addressed retrieval and scope isolationPASSPASSPASSPASSFAIL
D03Ranked relevance retrievalPASSFAILFAILPASSFAIL
D04Update and current valuePASSPASSPASSPASSFAIL
D05Lifecycle transition and removalPASSPASSPASSPASSFAIL
D06Logical scope isolationPASSPASSFAILPASSFAIL
D07Data-root isolationPASSPASSPASSFAILFAIL
D08Metadata filteringPASSN/AFAILPASSFAIL
D09Temporal boundaryPASSPASSPASSPASSFAIL
D10Revision historyPASSPASSFAILPASSFAIL
D11Relationship resolutionPASSFAILFAILFAILFAIL
D12Source tracePASSPASSPASSPASSFAIL
D13Batch persistence and exactly-oncePASSFAILPASSPASSFAIL
D14Restart recoveryPASSPASSPASSFAILFLAKY
D15Official exportPASSPASSPASSN/AN/A
D16Official importPASSPASSPASSN/AN/A
D17Concurrency, idempotency, integrityPASSFAILPASSFLAKYFAIL
D18Invalid-input containment and recoveryPASSPASSPASSPASSFAIL
Stable PASS / applicable dimensions (N/A excluded from the applicable denominator)
ProductLocked versionStable PASSFAILFLAKYN/A
Aethmere0.7.0 · 2c1df7118 / 18000
Letta Code0.30.2913 / 17401
MemOS2.0.3013 / 18500
Mem02.0.1812 / 16312
Graphiti0.29.31 / 161412

10.00 means only that Aethmere achieved 18 / 18 stable two-run passes in this locked V5 contract; it is not a perfect score for arbitrary product quality. Graphiti's 14 stable FAIL outcomes were native operations that did not finish within the locked limit, not proof that the functions are absent. A different operator ran this owner-run evaluation on an independent physical Apple Silicon Mac. It is not third-party certification, an open-world leaderboard, or a production SLA.

Download V13 results and formula ↓Methods, hashes, and limitations ↗
PUBLIC DOWNLOADS

The four downloads serve different jobs; you do not need all of them.

For Codex or Claude Code, install Agent Client only. Agent Studio is the standalone desktop chat app. The VS Code extension is optional, and the evaluation CLI only verifies public data. Every file is available directly from the official site with a public SHA-256 checksum.

RecommendedNODE · TGZ

Agent Client

Save project context locally and connect it to Codex or Claude Code through MCP. This is the only component required for everyday integration; it makes no network requests and has no telemetry.

npm install -g https://aethmere.com/downloads/aethmere-agent-0.10.0.tgz
Download TGZ package ↓GitHub Release ↗
aethmere-agent-0.10.0.tgz · 6.9 KB
Standalone desktop chatWINDOWS X64 · ZIP

Agent Studio

Portable Windows desktop app that requires Ollama to be installed and running locally. It does not replace Agent Client for Codex / Claude Code, does not scan projects automatically, and has no telemetry.

Fully extract the ZIP, then run Aethmere Agent Studio.exe
Download for Windows ↓Verify SHA-256 ↗
aethmere-agent-studio-0.11.0-win32-x64.zip · 143.2 MB · unsigned preview
OptionalVS CODE · VSIX

VS Code extension

Optional helper: save selected editor text to local context, inspect entries, and copy the MCP connection command for the current project. After downloading, choose “Install from VSIX” in the VS Code Extensions view.

Extensions → … → Install from VSIX
Download extension ↓GitHub Release ↗
aethmere-vscode-0.10.0.vsix · 13.0 KB
Evaluation dataNODE · TGZ

Evaluation CLI

Only for verifying the historical V3/V5 aggregate, running public evidence-ID examples, and checking release hashes across the website and GitHub. It does not connect Codex or Claude Code.

npm install -g https://aethmere.com/downloads/aethmere-cli-0.10.2.tgz
Download evaluation CLI ↓GitHub Release ↗
aethmere-cli-0.10.2.tgz · 5.9 KB
GOOD FIT

Built for work that takes more than one conversation.

If one conversation solves the problem, you do not need another layer. Long-running work that advances over many sessions is where saved context becomes useful.

01
Long-term research

Save research goals, decision evidence, and questions that still need confirmation.

02
Complex projects

Carry requirements, constraints, and key decisions across multiple sessions.

03
High-stakes decisions

Use stable IDs to point to context you can actually inspect and reduce unsupported completion.

YOUR CONTEXT, YOUR BOUNDARY

What stays local, and when does a component use the network?

The network boundary depends on the component: Agent Client is fully local; Agent Studio uses local Ollama by default and sends minimum conversation text only when cloud governance is configured.

01Agent Client is offline and has no telemetry; project context stays in .aethmere/context.json inside the current project
02Studio keeps conversations, model selection, and project selection local. Configured governance sends only minimum conversation text—never project paths, files, Git data, shell data, or tokens
03Every release file includes a version, size, and SHA-256 checksum
ABOUT AETHMERE-OS

Aethmere-OS is developed and operated by 青岛果冻先生网络科技有限责任公司.

We build Aethmere-OS (Shízhòu) around durable project context, verifiable memory, and clear local-data boundaries for long-running AI work. This official site provides product information, public downloads, evaluation methods, and a security-reporting path.

START LOCALLY

Start with one real project and let Codex or Claude Code read the context you chose to save.

View the 5-minute setup