工具上新 趋势信号

Hermes Agent:
8 周 10 万 Star 的开源 AI,凭什么?

它不是又一个聊天机器人,也不是写代码的助手。它是一个「住在你电脑上、越用越聪明」的私人 agent。最近中文圈讨论很多,值得了解一下它到底是什么。

📌 来源:Nous Research / GitHub / Hongkiat / DevPik · hermes-agent.org · 2026.05
10万+
GitHub Stars
295
社区贡献者(v0.13)
每周
发布一个新版本

大多数 AI 工具是这样的:你打开一个对话,聊完关掉,下次再打开它什么都不记得了。Hermes Agent 不一样。它装在你自己的电脑或服务器上,一直在后台跑着。你上次让它做了什么、你的偏好是什么、它自己学会了什么技能——这些全都记着,下次直接用。

一句话版本:它是一个本地运行、持久记忆、自我进化的私人 AI agent。由 Nous Research 开发(就是做 Hermes 系列模型的那个实验室),MIT 开源,免费。

它跟 Claude / ChatGPT 有什么不同

🧠
三层记忆
会话记忆(当前对话)、持久记忆(跨 session 保存,本地 SQLite)、自进化记忆(从经验中提炼出可复用的技能)。Claude 和 ChatGPT 也有记忆功能,但它们是云端的,Hermes 的全在你自己机器上。
🔄
自我进化的 Skill 系统
你让它做了一件事,它做完后会把过程写成一个「技能文件」存下来。下次遇到类似的事,直接调用这个技能,不用重新想。技能是 Markdown 文件,你可以看到、可以编辑、可以分享给别人。
🏠
永远在线
装好之后它作为系统服务在后台运行。你可以通过 Telegram、Discord、Slack、WhatsApp 给它发消息,它随时响应。不是你打开一个 app 才能用,而是它一直在那。
定时任务
内置 Cron 调度。你可以让它每天早上 8 点帮你汇总新闻,每周五下午生成周报,或者定时检查某个数据。不需要你手动触发。
🔒
数据完全本地
所有记忆、技能、对话历史存在你自己机器上的 SQLite 数据库里。没有遥测、没有跟踪、不经过任何第三方云服务。
🔌
40+ 内置工具
网页搜索、浏览器自动化、图片生成、TTS、代码执行,还支持 MCP 协议连接更多外部服务。模型可以接 Nous Portal、OpenRouter、OpenAI,或者你自己的本地模型。

最近火的是 Desktop 版

Hermes Agent 本身是命令行工具。最近小红书上讨论的是一个叫 Hermes Desktop 的项目——社区开发者 fathah 做的第三方桌面端,不是 Nous Research 官方出品,但把命令行里的操作搬到了图形界面里。

Desktop 版让人觉得「不只是换了个皮」的几个功能
Profile 切换 — 建多个身份,各自独立的配置、模型、人设、工具开关,互不干扰
Cron 可视化 — 以前定时任务是黑箱,现在能直观看到有哪些任务在跑
记忆可编辑 — 以前查记忆敲命令看纯文本,现在在界面上直接改,还能看容量
3D 办公室 — 内置一个叫 Hermes Office 的三维空间,看 agent 在「办公室」里的状态。花哨,但确实新鲜

⚠️ Desktop 版是第三方社区项目,稳定性还没经过大规模验证。如果你要试,先从 Hermes Agent CLI 本体开始,确认好用了再加 Desktop 层。

跟你已经在用的工具怎么比

Hermes AgentClaude CodeChatGPT / Codex
定位always-on 私人 agentsession 制编程助手通用 AI + 编程
运行方式后台常驻开 session 用完关云端/本地 session
记忆三层,全本地CLAUDE.md + 自动记忆云端记忆
自进化自动写 skill 文件不自动,你手写Custom GPTs
消息触达Telegram/Discord/Slack/WhatsApp终端 + 手机遥控App + 手机
数据位置你的机器你的机器(代码),云端(对话)云端
上手门槛一行命令安装,但配置多一行命令安装开浏览器就用

适合谁,不适合谁

如果你想要一个「一直在后台帮你盯着事情」的 agent——汇总消息、定时跑任务、记住你的偏好、越用越懂你——Hermes 是目前最认真在做这件事的开源项目。

如果你只是需要一个帮你写代码、改文章、做分析的工具,Claude Code 和 ChatGPT 已经够了,没必要为了 Hermes 多维护一套系统。

目前它还在 v0.13,不是 1.0。每周发版说明迭代快,也说明还在频繁改动。想尝鲜的可以装,想稳定用的再等等。

为什么值得关注

不是因为 10万+ star。是因为它代表的方向:AI 从「你开一个对话用完关掉」变成「一个一直在你身边、越来越懂你的助手」。这个方向不管最后赢的是 Hermes 还是别的什么,都是 AI 工具接下来要走的路。

Claude 在做记忆和 Projects,ChatGPT 在做 Custom GPTs 和 Memory,Hermes 用开源和本地化走了最激进的一条路。三家路线不同,终点可能相似。

TOOL UPDATE TREND SIGNAL

Hermes Agent:
Why This Open-Source AI Picked Up 100K Stars in 8 Weeks

Hermes is not another chatbot or another coding helper. It is a private agent that lives on your machine and gets more useful as you use it.

Source: Nous Research / GitHub / Hongkiat / DevPik / hermes-agent.org / 2026.05
100K+
GitHub stars
295
community contributors in v0.13
Weekly
new release cadence

Most AI tools work like this: you open a chat, finish the conversation, close it, and next time it remembers very little. Hermes Agent is built differently. It runs on your own computer or server, keeps working in the background, and remembers what you asked it to do, what you prefer, and what skills it learned along the way.

The short version: Hermes Agent is a local-first, persistent-memory, self-improving private AI agent. It is developed by Nous Research, the lab behind the Hermes model family, and released under MIT open source.

How It Differs From Claude or ChatGPT

1
Three layers of memory
Hermes keeps session memory, persistent memory across sessions in local SQLite, and self-evolving memory that turns repeated experience into reusable skills. Claude and ChatGPT also have memory features, but Hermes keeps them on your own machine.
2
Self-improving skill system
After it completes a task, Hermes can write the process into a skill file. The next time a similar task appears, it can reuse that skill instead of reasoning from scratch. Skills are Markdown files you can inspect, edit, and share.
3
Always online
Once installed, Hermes runs as a background service. You can message it through Telegram, Discord, Slack, or WhatsApp. It is not something you open only when you need a chat; it is meant to be present.
4
Scheduled tasks
Cron scheduling is built in. You can ask it to summarize news every morning, generate a weekly report every Friday, or check a dataset on a schedule without manually triggering the task.
5
Local data by default
Memory, skills, and conversation history live in a SQLite database on your own machine. The project emphasizes no telemetry, no tracking, and no third-party cloud service in the default path.
6
40+ built-in tools
It includes web search, browser automation, image generation, TTS, code execution, and MCP support for connecting more services. Models can come from Nous Portal, OpenRouter, OpenAI, or your own local setup.

The Recent Buzz Is About the Desktop Version

Hermes Agent itself is a command-line tool. The version being discussed recently is Hermes Desktop, a third-party desktop interface by the community developer fathah. It is not an official Nous Research product, but it brings many command-line operations into a graphical interface.

What makes the Desktop layer feel more than a skin
Profile switching - separate identities with their own settings, models, personas, and tool permissions.
Visual Cron - scheduled tasks become visible instead of hidden behind commands.
Editable memory - inspect and edit memory directly in the interface, including capacity and stored items.
3D office - a visual Hermes Office space that shows the agent's status. It is flashy, but it is also genuinely new.

The Desktop version is a third-party community project and has not been validated at large scale. If you want to try Hermes, start with the core CLI first, then add the desktop layer if it proves useful.

How It Compares With Tools You May Already Use

Hermes AgentClaude CodeChatGPT / Codex
Positioningalways-on private agentsession-based coding assistantgeneral AI + coding
Runtimebackground serviceopen a session, close it latercloud/local sessions
Memorythree layers, localCLAUDE.md + auto memorycloud memory
Self-improvementwrites skill filesmanual configurationCustom GPTs
MessagingTelegram/Discord/Slack/WhatsAppterminal + mobile remote controlapp + phone
Data locationyour machineyour machine for code, cloud for chatcloud
Learning curveone-line install, many settingsone-line installopen browser and use

Who It Is For, and Who Can Skip It

If you want an agent that keeps watching things in the background - summarizing messages, running scheduled tasks, remembering preferences, and learning from repeated work - Hermes is one of the most serious open-source attempts right now.

If you only need help writing code, editing documents, or doing analysis, Claude Code and ChatGPT are already enough. There is no need to maintain another system just for the sake of it.

It is still at v0.13, not 1.0. Weekly releases show fast momentum, but also frequent change. It is good for experimenting, less ideal if you need a stable daily system today.

Why It Is Worth Watching

Not because of the 100K stars. Because it represents a direction: AI tools are moving from "open a chat, use it, close it" toward "an assistant that stays with you and gets to know your workflow." Whether Hermes wins or not, this direction matters.

Claude is building memory and Projects. ChatGPT has Custom GPTs and Memory. Hermes takes the most aggressive local-first and open-source route. The paths differ, but the endpoint may rhyme.