Aeon vs Grok Bot: the $300 agent and the free one

Aeon vs the Others, #4. Grok Bot rents you a fleet of cloud agents. Aeon gives you one you own.

This is the closest fight in the series. Both are always-on agents that live in the cloud, run on schedules, and keep working with your laptop shut. Grok Bot is xAI's beta (built with Cursor, launched August 11, 2026): named agents, each with its own cloud computer, that sign into your apps and drive the screen like an employee. Aeon is an open-source framework you fork: skills are Markdown files running on GitHub Actions, scored on every run, patched when they break.

Same promise, work that happens without you. Opposite answers on who owns the agent, where your credentials live, and what it costs. And there's a twist: Grok is one of Aeon's ten harnesses, so Aeon can run on Grok for a lot less than Grok Bot does.

Set up Aeon. Free, MIT-licensed, and it runs on your own GitHub account.

Aeon is a framework you fork and run. A skill is a Markdown file with frontmatter and a prompt, scheduled on cron or fired by reactive triggers, executed on GitHub Actions by any of ten harnesses (Claude, Grok, Codex, Pi, Vibe, Kimi, fx, Cursor, Hermes, GLM). A model scores every run 1-5, and a self-healing loop (heartbeat, skill-health, skill-repair) diagnoses and patches skills that fail. It reports to Telegram, Discord, Slack, Buzz, or email, and can spawn specialized copies of itself into a fleet.

Grok Bot is xAI's agent platform, in beta since August 11, 2026. You create named bots (research bot, email triage, chief of staff), each with a dedicated cloud computer (browser, terminal, files). Bots log into your tools with your credentials, learn tasks by watching you do them once, run on scheduled "routines" or event triggers, and coordinate with each other in group chats, with a chief-of-staff bot delegating to specialists. Desktop and iOS, enterprise on waitlist.

• Runs without you - Aeon: GitHub Actions, cloud-native. Grok Bot: a dedicated cloud computer per bot.
• Scheduling - Aeon: native cron per skill. Grok Bot: routines, timed or event-triggered.
• Reactive triggers - Aeon: condition-based. Grok Bot: event-based (new message, commit).
• Multi-agent - Aeon: spawn-instance fleet, isolated billing. Grok Bot: chief-of-staff plus specialist bots.
• Self-healing - Aeon: auto-diagnoses and patches failing skills. Grok Bot: approval checkpoints, no repair loop.
• Quality scoring - Aeon: every run scored 1-5 by a model. Grok Bot: none.
• Model choice - Aeon: 10 harnesses, swap in config. Grok Bot: Grok only.
• Open source - Aeon: MIT, fork it. Grok Bot: closed, xAI plus Cursor infrastructure.
• Where credentials live - Aeon: sealed GitHub Actions secrets. Grok Bot: your app logins, stored in their cloud.
• Audit trail - Aeon: every run is an Actions log in your repo. Grok Bot: audit logs inside the product.
• Persistent memory - Aeon: versioned files (SOUL.md, STRATEGY.md). Grok Bot: remembers conversations, opaque.
• Drives a GUI (clicks, types) - Aeon: no, CLI and API work. Grok Bot: yes, operates real apps on screen.
• Teach by demonstration - Aeon: no, you write a Markdown skill. Grok Bot: yes, watches you do it once.
• Zero setup - Aeon: no, fork a repo and add secrets. Grok Bot: yes, name a bot and hand it work.
• Status - Aeon: shipping, self-hosted. Grok Bot: beta, week-to-week changes.
• Entry price - Aeon: free. Grok Bot: $120-$300/mo tiers.

Aeon: free, MIT-licensed. Actions minutes cost nothing on public repos. You bring model auth: a Claude subscription, or API keys for whichever harness you pick, including xAI's Grok API at normal token rates.

Grok Bot: no standalone plan. Beta access is bundled into SuperGrok Heavy ($300/mo), Cursor Ultra ($200/mo), or Cursor Teams Premium ($120/user/mo).

That gap is the headline. The cheapest door into Grok Bot is $120 a month. Aeon running skills through the Grok harness costs you Grok API tokens and nothing else.

• The work lives in GUIs: apps with no API, web dashboards, forms. Screen-driving is Grok Bot's party trick and Aeon doesn't do it.
• You want to teach by showing, not by writing. Demonstrate a task once, it becomes a routine.
• You're non-technical and allergic to repos, secrets, and YAML.
• You're already paying for SuperGrok Heavy or Cursor Ultra, so the marginal cost is zero.

• You want to own the agent: skills, memory, logs, all versioned files in a repo you control, MIT-licensed.
• Handing an AI your live app credentials in someone else's cloud makes you flinch. Aeon's secrets are sealed Actions secrets, and read-only skills can't write at all.
• You want the runtime to notice failures and repair them, not wait for you to check on a bot.
• You want model choice, or a hedge. Seven harnesses, swappable in config, Grok included.
• You'd rather pay ~$0 than $120-300/mo for scheduled autonomous work.

In one line: Grok Bot is an employee you rent, working inside xAI's building. Aeon is an agent you own, working inside yours.

Aren't these the same product?

Closest pair yet: both cloud-native, scheduled, multi-agent, built for unattended work. The split is ownership and mechanism. Grok Bot is a closed platform driving GUIs with your logins on xAI/Cursor infrastructure. Aeon is an open framework running Markdown skills on GitHub Actions in your own account, with scoring and self-repair that Grok Bot doesn't have.

Can Aeon really use Grok?

Yes. Grok is one of Aeon's ten harnesses. Point the config at it and your skills run on Grok models, paying API token rates instead of the $300/mo SuperGrok Heavy gate. You lose the screen-driving and the teach-by-demo, which are Grok Bot product features, not model features.

What about the credential question?

Grok Bot's whole model is signing into your Gmail, Slack, and CRM with your credentials from its cloud computer. Powerful, and a real attack surface, which even sympathetic coverage flags. Aeon never sees your app passwords: secrets are sealed as Actions secrets, never touch the command line, and irreversible actions fail closed.

Grok Bot has a chief-of-staff bot. Does Aeon have an equivalent?

Aeon's version is the fleet: spawn-instance creates specialized copies (research, trading, repo triage) with isolated billing and no shared secrets, and fleet-control coordinates them. Same idea, delegation to specialists, but each specialist is a repo you can open and read.

Grok Bot is in beta. Does that matter?

Some. Access is gated, platforms are still rolling out, and features change week to week, which is normal for a beta. Aeon is smaller and less flashy but shipping: fork it today, skills run today, and when one breaks the repair loop patches it.

Can I run both?

Sure. A sane split: Grok Bot for GUI-bound tasks no API can reach, Aeon for everything scriptable, scheduled, and worth auditing. If the monthly bill starts to sting, the scriptable half moves to Aeon for free.

• #1 - Aeon vs Hermes: two autonomous agents, one big difference
• #2 - Aeon vs Claude Code: same engine, nobody at the wheel
• #3 - Aeon vs ChatGPT: the agent that works while you're gone

Point any coding agent at https://www.aeon.fun/skills/aeon.md, or fork aeonfun/aeon and run ./aeon.

Configure once, forget forever!