Grok Bot vs Hermes Agent: which should you use?
Pick Grok Bot if you want AI teammates that work on day one with nothing to host: install the app, and xAI runs the computer and chooses the model, from $30 a month on The X Subscription. Pick Hermes Agent if you want to choose the model, keep the agent on your own machine or server, and pay only for what it uses: it is free and MIT-licensed, and its Bot Mode gives you the same roster of named bots, each with its own model, memory and routines. Hermes can even run Grok models through a SuperGrok or X Premium+ sign-in.
Grok Bot and Hermes Agent side by side
| Grok Bot | Hermes Agent | |
|---|---|---|
| What it is | xAI's managed AI teammates, built with Cursor | Nous Research's open-source agent with memory and a learning loop |
| License | Proprietary | MIT, about 254,000 GitHub stars |
| Several named bots | Yes: bots with names, jobs and memory | Yes: Bot Mode, where each Bot is a Hermes profile with its own model, memory, skills and routines |
| Where it runs | One managed cloud computer per account, in the US; each bot gets its own screen | Your computer, a VPS or Hermes Cloud. Bot Screen gives each bot its own desktop on a Linux server |
| Reach it from | Apps for macOS, Windows, Linux, iOS and Android, and @bot on X | Terminal, Hermes Desktop and 20+ chat apps: Telegram, WhatsApp, Slack, Signal, Teams, iMessage via BlueBubbles |
| Model choice | None: no model picker; Grok Bot picks a model per task | Any: 200+ models, a custom OpenAI-compatible endpoint, and a different model per Bot |
| Price | X Subscription Premium $30, Super $100 or Ultra $200 a month, or a paid Cursor plan | Free. You pay for the models it calls, from $0 on a local model |
| Usage limit | Weekly pool shared with Grok and Cursor | Whatever your model provider or plan sets |
| Training on your data | Kept out of training only with Privacy Mode on | Hermes is self-hosted; your model provider's policy applies |
| Safety rails | Auto Review of risky actions; network allowlists on Enterprise | Command approvals and an optional Docker or SSH sandbox; you own the setup |
Sources: Grok Bot overview, security, models, plans and X Subscription pricing; Hermes Bot Mode, Bot Screen, messaging and providers docs, and the repo. Checked October 15, 2026.
What Grok Bot is
xAI describes Grok Bot as "AI teammates with names, jobs, and context that compounds over time" (overview). Every bot on your account shares one cloud computer with a browser, files and a terminal, each on its own screen, and keeps working while your laptop is closed. You reach them from apps on macOS, Windows, Linux, iOS and Android, or by tagging @bot in a post on X. They use connectors where they exist and computer use for everything else.
It comes with The X Subscription's Premium ($30 a month), Super ($100) and Ultra ($200) tiers, with bot usage drawn from a weekly pool shared with Grok and Cursor; Plus ($8) does not include it (pricing). Paid Cursor plans include it too, and SuperGrok, SuperGrok Heavy and X Premium+ subscribers get it by linking their account (plans). The free trial is a usage credit with a seven-day window.
You can't choose the model. Cursor's help page says "Grok Bot has no per-user or per-request model selection, and there is no model picker": it picks an xAI or third-party model per task (models), and there is no API key or base URL setting.
What Hermes Agent is
Hermes Agent is Nous Research's MIT-licensed agent: it remembers you across sessions, writes reusable skills from finished tasks, runs scheduled jobs, and answers in the terminal, Hermes Desktop or more than 20 chat apps (repo).
The closest thing to Grok Bot is Bot Mode, on by default in Hermes Desktop: a roster of named Bots, each a Hermes profile with its own role, model, memory, skills and routines, which message each other and share group chats (Bot Mode). On a Linux server or Hermes Cloud, Bot Screen gives each bot its own desktop that you can watch live and take over for a login or 2FA prompt, and the bot keeps working with your laptop off (Bot Screen).
Hermes is free. The models are the bill: Nous Portal credits, a ChatGPT sign-in, your own API key, a local model or a fixed-price API. It even signs in with SuperGrok or X Premium+ to use Grok models, though xAI returns HTTP 403 on some tiers, in which case Hermes's docs say to use an xAI API key instead (providers).
Where Grok Bot is better
- No setup. You install an app and the bots have a computer. Hermes needs a machine that stays on, a VPS or Hermes Cloud, and someone to update it.
- One price. Premium at $30 covers the computer, the model and the apps. With Hermes you pick and pay for the model, and a frontier model on a pay-per-token key can cost more than $30 in a busy month.
- Managed safety rails: Auto Review checks risky actions before they run, and Enterprise admins set network allowlists. Hermes's Bot Screen docs say screens are work surfaces, not security boundaries, so isolation is what you configure.
- X built in: tag @bot on a post and it becomes a task.
Where Hermes Agent is better
- Model choice. Each Bot can run a different model, so a scheduled inbox summary uses a cheap model and a hard research task uses a frontier one. Grok Bot decides for you.
- Cost control. A free agent plus a model bill you set: $0 on a local model, about $6 a month for an always-on agent on an efficient model at list prices, or a fixed plan that stops at its budget.
- Where your data goes. Grok Bot computers run in the US and data is kept out of training only with Privacy Mode on. Hermes runs where you put it and sends prompts only to the model API you choose.
- Chat apps. Hermes answers in Telegram, WhatsApp, Slack, Signal, Teams, iMessage and more. Grok Bot lives in its own apps and on X.
- No lock-in. MIT-licensed, about 254,000 GitHub stars, with memory and skills stored as files you own.
Run Hermes Agent on one API
A roster of bots that run routines all day makes far more model calls than a chat. One OpenAI-compatible endpoint keeps it to one key and one bill.
- Cost first: send the model
standardcomputeand the router sends each step to the cheapest model that can do it. Routine steps such as reading pages, filling forms, summarizing and scheduled check-ins can go to low-cost models, and Claude Fable 5.1 or GPT-6 Astra handle the steps that need them. An exact model ID, such asanthropic/claude-sonnet-5.5, always stays on that model. - For every model we vet providers for price, uptime and speed, with automatic fallback when one fails, so a bot that runs overnight does not stop because one host is down.
- Opt-in Compliance rules restrict a key, team or company to provider endpoints in Europe or the US. They cover where the model runs. The bot's computer runs wherever you host the project.
- We don't train on your prompts or store them.
- Plans start at $19 a month for a fixed compute budget. At the end of the budget requests stop with HTTP 402 until renewal, so a looping bot cannot run up a bill.
Set up Hermes Agent on Standard Compute
- Install Hermes Agent, then the Standard Compute launcher, and run
standardcompute launch hermes. It saves a Standard Compute profile and leaves your usual Hermes profile alone. - Without the launcher: run
hermes model, choose Custom endpoint, and enterhttps://api.stdcmpt.com/v1, your Standard Compute key and the modelstandardcompute. - In Bot Mode, leave a Bot's model pin empty to inherit that model, or pin an exact ID such as
anthropic/claude-sonnet-5.5on the one Bot that needs it.
The catch
Hermes is a project you run, not an app you install: you keep a machine on, update it and decide what each bot may touch, and a bot with your browser and accounts can do real damage if a prompt injection lands. If you won't do that, Grok Bot is the better pick, as long as you accept xAI choosing the model and a US-hosted computer.
Where Standard Compute fits
One OpenAI- and Anthropic-compatible key for every Hermes Bot, with a router that sends each step to the cheapest model that can do it, providers vetted for price, uptime and speed, opt-in Europe or US rules and no training on prompts. Plans start at $19 a month.
FAQ
Is Hermes Agent better than Grok Bot?
For control and cost, yes: Hermes is free, runs on any model you choose and keeps your data where you host it. For zero setup, no: Grok Bot comes with a managed computer, apps on every platform and safety review, and you never touch a server.
Does Hermes Agent have bots like Grok Bot?
Yes. Bot Mode, on by default in Hermes Desktop, turns Hermes profiles into a roster of named Bots, each with its own model, memory, skills and routines. Bots message each other and share group chats, and on a Linux server each one can get its own desktop through Bot Screen.
Can Hermes Agent use Grok models?
Yes. Choose xAI Grok OAuth (SuperGrok / Premium+) in hermes model to sign in with a SuperGrok or X Premium+ subscription, or use an xAI API key. Hermes's docs note that xAI returns HTTP 403 on some subscription tiers, and the fix is an API key.
Can Grok Bot use my own model or API key?
No. Cursor's help page says Grok Bot has no per-user or per-request model selection and no model picker, and there is no API key or base URL setting.
Which is cheaper, Grok Bot or Hermes Agent?
Hermes, for most people: the agent is free, and an always-on agent on an efficient model costs about $6 a month at list prices. Grok Bot starts at $30 a month on X Subscription Premium. Hermes on a pay-per-token frontier model can cost more, which a fixed-price plan avoids.
Is Grok Bot free?
No. It has a free trial, a usage credit with a seven-day window, and then needs X Subscription Premium ($30), Super ($100) or Ultra ($200), a paid Cursor plan, or a linked SuperGrok or X Premium+ account.