Flagship product · technology

The content factory: how AI runs the newspapers of whole cities

Direction No. 2 of the MILAGPT cooperative. While the market debates whether AI can be trusted with a single post, our AI agents run the newspapers of two cities every night. I'll tell you what this is, why it's a window, and where we're heading.

July 8, 2026 · MILAGPT Blog

Two factories, not one

MILAGPT has two engines. Direction No. 1 is the Telegram bot factory: we produce bots and nanobots in series. Direction No. 2, the subject of this article, is the content factory.

They don't compete; they feed each other. The bot factory makes the product. The content factory makes the stream of attention that sells that product. One without the other is half a business.

The essence of direction No. 2 in one sentence: point AI, in autonomous mode, at gathering, analysing and preparing material — and use that content to run media automatically.

What a content factory really is

"AI and content" usually means: a person sits down, asks a model to write a post, edits it, publishes it. That isn't a factory. That's sped-up manual labour.

A factory is when the loop is closed with no human inside. The AI gathers data on a city and a topic itself. Analyses it itself. Prepares the material itself. And that content itself runs two types of platforms: websites in the form of city newspapers, and social media — Telegram channels, Instagram, YouTube, Twitter, LinkedIn.

An important nuance that often gets lost: gathering and analysis aren't the goal in themselves. They exist for the sake of the final product — quality content. Data without an edition is dead. We build the whole pipeline, from raw material to distribution.

The city newspaper is our own development

The key format of the content factory is the city newspaper. It's not a metaphor and not an ordinary news site. It's a living graph of the city that works as a media outlet.

Take ToshkentToday. Behind it stands a graph of about 6,900 organisations in Tashkent. Every night AI agents update it: they gather, cross-check, rewrite, publish. In the morning the city gets a fresh newspaper, assembled overnight not by a news desk but by a pipeline.

This is hard to copy, and that's the whole point. You can copy a website's design. You can't copy the gathering and analysis infrastructure that fills it every night. The newspaper is the shop window, and underneath it is the engine of a city aggregator.

You can copy a website's design. You can't copy the pipeline that fills it anew every night.— Shakhruz Ashirov, founder of MILAGPT

It already works. In two cities

We're not selling a slide with a promise. We're showing two live cities.

ToshkentToday — Tashkent, our first and anchor city. NorthCyprusToday — the second city, and importantly, it was launched by an outside partner. That means the model has already been reproduced by other hands in another jurisdiction. That's the main test: does it work only for us — or as a system?

The answer: as a system. And since the system reproduces, it can be scaled not by hiring but by cloning the pipeline to a new city.

A network of cities, not a franchise

Here there's a fundamental fork, and we chose our side deliberately. We're building not a franchise but a network of cities.

A franchise is when a partner rents the brand and pays royalties while staying outside. A network is when a partner comes inside and becomes a co-owner. Whoever launches a new city isn't a tenant of our name but a member of the cooperative with a stake.

That changes everything. A tenant's motivation is to recoup the payment. A co-owner's motivation is to grow the shared network, because their stake grows. Through the MILA token we turn every new city into a joint asset, not a one-off deal. That's our moat: a competitor has to build not only the technology but also a network of motivated co-owners. The second is harder than the first.

How it makes money

Content isn't charity. It earns, and in several layers at once.

The first layer is selling our bots and factories. The city's newspaper and social media create reach, and reach sells the products of direction No. 1: bots, nanobots, whole factories for partners joining the network. Direction No. 2 is essentially a traffic machine for direction No. 1.

The second layer is cooperative services for local business: turnkey social media management, analytics, lead generation straight from the city graph, priority placement and advertising inside the newspaper. The third layer is the city's share, which goes to the cooperative's dividend pool, and the MILA token as an instrument of co-ownership. One pipeline — several sources of income per city.

The window is closing — we're going in first

Autonomous content factories are a window of opportunity, not a permanent head start. Today only a handful can point AI at the full cycle of gathering → analysis → edition → distribution. In a year or two this will be the norm, and there will be no one left to stake out cities — they'll be taken.

So our strategy is simple and aggressive: go in first, city by city, while the market is still arguing whether AI can be trusted with a post. We've already moved past that argument — we have two live cities and a reproducible model.

Where the cooperative is heading: from two cities to a network of cities where each node runs itself, feeds traffic to the bot factory, serves local business and brings a share into the common pool. We're not building a website or a newsroom. We're building infrastructure that produces the media of whole cities on autopilot.

This is an article from the MILAGPT "Blog" section. Here we take apart the ideas we're building the cooperative on — and sometimes share news. · ← All articles · Home · White Paper