Can ChatGPT Dots Replace Hermes Agent?
Dots is ready to use, Hermes Agent you set up yourself. Which fits, judged by how complex your needs are, from Google Drive sync to WhatsApp, plus the cost.

Yes, for some people. If you've been using Hermes Agent as a personal assistant to read email and Google Drive, remind you of your schedule, and prepare documents, and you're getting tired of looking after the server, ChatGPT Dots can replace it. But if you need to chat through WhatsApp or Telegram, choose your own model, keep data on your own server, or plug in custom tools, Dots can't replace Hermes yet.
What decides it isn't which one is smarter, but how complex your needs are and how ready you are to handle the technical side.
Two different approaches
ChatGPT Dots is OpenAI's ready-to-use agent. According to the Dots help page, a dot is powered by GPT-6 Astra, has its own cloud computer, brings context from your ChatGPT memory, and connects to apps through plugins. Your first dot is included in the ChatGPT Pro plan and Business Premium seats.
Hermes Agent is an open source agent from Nous Research under the MIT license. According to its official GitHub repository, Hermes can be installed on Linux, macOS, WSL2, Windows, and even Android through Termux, and can use any model through Nous Portal, OpenRouter, OpenAI, or your own endpoint. Hermes has a gateway to Telegram, Discord, Slack, WhatsApp, Signal, and email, a built-in cron scheduler, memory across sessions, skills it creates from experience, and MCP support for adding tools.
I introduced Hermes and similar platforms in Agentic AI: When You Actually Need It, and Platforms You Can Run Directly on a VPS.
Compare by need
| Need | ChatGPT Dots | Hermes Agent | Easier with |
|---|---|---|---|
| Reading and working with Google Drive files | Connect the Google Drive plugin, sign in, approve permissions | Install the Google Workspace skill, set up a Google Cloud project and OAuth credentials | Dots |
| Chatting through WhatsApp or Telegram | Not available. The options are the ChatGPT app, phone calls, Slack, and texting in beta in the US | Yes, through one gateway process | Hermes |
| Scheduled tasks and reminders | Yes | Yes, a built-in cron that can deliver results to any platform | Same |
| Choosing your own model, including local models | No, it uses OpenAI's models | Yes | Hermes |
| Keeping data on your own server | No, it runs in OpenAI's cloud | Yes, it runs on your laptop or VPS | Hermes |
| Maintenance, updates, and server security | Handled by OpenAI | You handle it | Dots |
| Adding custom tools or integrations | Through the available plugins | Through skills and any MCP server | Hermes |
The clearest example, syncing with Google Drive
In Dots, Google Drive connects through the same Plugins menu as ChatGPT. According to the Google Drive guide for ChatGPT, you pick Google Drive, sign in to your Google account, review the permissions, and complete the authorization. After that, ChatGPT can search files, read Docs, Sheets, and Slides, and run actions such as creating or updating files if you grant the permission. Two notes: actions that change a file may ask for confirmation first, and the apps available to a dot may differ from regular ChatGPT.
In Hermes, the route is the Google Workspace skill, which covers Gmail, Calendar, Drive, Contacts, Sheets, and Docs. The steps are more technical:
- Create a Google Cloud project, then enable the Gmail, Calendar, Drive, Sheets, Docs, and People APIs.
- Create OAuth 2.0 credentials of the Desktop app type, then download the JSON file.
- Open the authorization link Hermes generates, approve it, then paste the redirect URL back.
- After that, the token refreshes automatically.
Hermes walks you through this, but you still need to be comfortable in the Google Cloud Console. If you only need email, the documentation suggests the himalaya skill, which works with a Gmail App Password and no Google Cloud project.
So for someone who just wants the agent to read Drive, Dots is much simpler. For someone who wants the agent to work across Gmail, Calendar, and Drive at once and report back on Telegram, Hermes is more flexible.

The cost
Illustrative monthly cost(thousand rupiah)
- ChatGPT Pro (Dots)Rp1.889M
- Hermes on a VPS + Nous Portal Plus±Rp499k
Hermes itself is free, but two costs keep running: the server and the model. On the official Hermes site, Nous Portal plans for models are available free, at $20, $100, and $200 a month, and you can also use your own API keys. I cover VPS price benchmarks in Indonesia in What Is a VPS for AI, and What Can It Actually Do?.
On paper, Hermes is cheaper (roughly US$28 a month in this example, against ChatGPT Pro's $100). But count the time you spend installing it, updating it, and fixing it when something breaks. For someone whose time is expensive, that price gap can disappear into working hours.
Who each one suits
- Non-technical owners or professionals who already use ChatGPT: Dots. Ready to use, no server.
- Teams in Indonesia that work through WhatsApp or Telegram, with someone who understands servers: Hermes.
- Developers who want cheap or local models and full control of their data: Hermes.
- Companies that need admin controls: Dots through Business Premium seats or Enterprise, or Hermes on a company server with security they design themselves.
If you don't use either yet, I'll compare Dots with other managed agents, such as Grok Bot and Meta Muse, in the next article.
What people often forget when installing Hermes
Hermes can run commands and access files on the machine it's installed on. That's what makes it powerful, and also what makes it risky. Install it on a VPS or dedicated device that doesn't hold other sensitive data, give it the least account access possible, and update it regularly. With Dots, OpenAI takes over the server side, but who holds the agent's access and memory still needs managing, as I cover in ChatGPT Now Has Dots, and What's Interesting Isn't How Smart It Is but Who Holds the Access.
Wrapping up
Dots can replace Hermes for simple needs and for people who don't want to look after a server. For more complex needs, like WhatsApp, a model of your choice, data on your own server, or custom integrations, Hermes is still ahead. Choose based on the most complex need you actually have, not features you might use someday.
If you'd like Hermes Agent or a similar agent installed and secured on your own server, or just want to know which fits your team better, tell me what you need.
The photos in this article are from Unsplash, by Daniil Komov (main photo) and Ilnur, used under the Unsplash License.
Is your team still doing something by hand?
Tell me about it, and let's find the fix together.
More notes

Can AI Learn Your Sales Style From WhatsApp Chat History?
Yes, and you usually don't need to train your own model. How to export WhatsApp chats, turn them into a style guide, a ready-to-use prompt, and the customer data risks.
Read the note
Gemini vs ChatGPT for Office Work, Which One Fits Better?
Gemini and ChatGPT can both be used for work, but their strengths differ. A plan price comparison in Rupiah, which suits teams on Google Workspace, and when ChatGPT makes more sense.
Read the note