Six GrokBots Form a City; He Says the Content Runs Itself While He Sleeps
Morlex breaks content production into six workstations—finding topics, testing openings, replying to comments, rendering videos, publishing, and doing the books—and turns the interface into a city. Someone in the comments asked a very practical question.
Morlex posted on X with a video of a stroll through a city. Six GrokBots, one city, and while he sleeps, the content pipeline runs on its own.
This setup is called Grok City. He breaks content production into six workstations, each with one bot:
- Scout stays in the Newsroom and is responsible for finding an AI news story before it goes viral
- Hook is in the Writers' Room, testing three openings against Grok; Morlex picks the winner with one click
- Reply watches over the Café, answers comments, and drafts sponsorship deals along the way for him to approve
- Reel renders videos in the Studio and sends them to the Post Office when done
- Post publishes the thread and video to X, and a paper plane flies out in the system
- Ledger sits in the Bank and tracks how much each post earns
A content pipeline tracks every task, from story to payment. A real-time activity log shows who is doing what. According to Morlex, he makes only four decisions across the entire chain; the rest is left to the city.
The video shows the whole city: on the left, the task bar lists needs you, approvals, and the timeline; on the right is the pipeline status from found to counted; today's stats hang in the corner—8.2k views, 12 new followers, and $120 in revenue. The interface looks like a business simulation game.
Want to build one yourself? In the replies, Morlex gave a starting template for each bot:
- scout starts with Researchy
- hook starts with Copy Humanizer
- reply starts with Event Request Desk
- reel starts with Director
- post uses any bot plus the Postiz plugin
- ledger creates a new bot connected to read-only Stripe
The links are all under the original tweet.
The comments are more interesting than the tweet itself.
Someone asked directly: How do you know it actually used that model? In the screenshot, Grok Bot's answer is that it cannot disclose which models run underneath or switch them; all it can do is research, write code, get work done through cloud agents, manage calendar and email connectors, and then hand the work off to other bots. Another user, Ingeniero, gave shorter feedback: Not recommended; it burns through tokens too fast.
These two pieces of feedback point to two concrete problems. On one hand, the model is opaque: you cannot verify from the outside who fired the shot. On the other, cost: with six bots online at once, the bill won't get gentler just because the interface looks like a city. Others care about something more basic: how was this visualized city built?
What's really worth watching in this approach may not be 'passive income while you sleep.' Breaking content production into replaceable workstations, each with an off-the-shelf template, clear inputs and outputs, and an activity log—that idea has reference value for people building content pipelines. As for whether it actually makes money while you sleep, you'll have to look at next month's bill and backend data.
Morlex says the full playbook is in the article. The city is built; visitors are welcome.
发布时间: 2026-11-01 18:48