How Content OS finds and posts each story

Shopping for a social media scheduling tool? This is what actually happens on every run today, written out so you can check it before you commit, followed by what is designed and not built yet.

What does Content OS actually do?

It finds the story, writes it fresh, renders it and posts it to your own pages, on the schedule you set. Four steps, every run, in this order.

1. It finds the story. Each run picks today’s story from trends and the material available, and drops the thin candidates rather than padding them out.
2. It writes it fresh. The post is written new for this run, so the page is not cycling a bank of old captions.
3. It renders it. The words go into a format that declares its text limits, and the finished image or video is rendered from that format.
4. It posts it. The finished post goes out to your own pages, through your own signed in browser session.

Most tools stop well short of that. A scheduler hands you an empty calendar and asks what goes in it. An AI writer hands you a pile of drafts and asks which one is good. Both leave the finding and the making where they were.

How do I set it up?

Three steps, and the first two happen once. There is no approval queue to clear afterwards, because there is no approval step: the run posts on its own.

1. Connect your own pages. Sign in to your own pages in your own browser, on the machine that posts. You never hand over a password.
2. Set the schedule. Give each scheduled job the times it should run. You can pause any job, or press Run now to start a run straight away.
3. Let the run go. Each run finds today’s story, writes it fresh, renders it into a format and posts it to your own pages.

That is the whole of the control surface, and we would rather say so plainly than dress it up. A job you want to hold is paused. A run you want now is one press away. Everything else happens on the schedule.

Which story does it pick?

Today’s, from what is moving and what material there is to back it. Each run looks at trends and at how much material is available for each candidate, and a candidate with too little behind it is dropped.

That is the picking that runs today. It does not weigh a list of topics against each posting time, and it does not keep a scored record of what won and what lost. That fuller planner is designed and described under what is coming below, and we would rather you read it as future than assume it is live.

What happens when nothing is good enough?

The thin candidates are dropped rather than posted as filler. We would rather the page skip a weak story than post one.

That sounds like a small thing and it is not. A scheduler with nothing in it is a failure you have to notice; a run that drops a thin story has refused to pad the page. Every run writes a log with its status and how long it took, and the admin console shows it, so a run that did not post is visible rather than silent. If thin days keep appearing, that is the page telling you it needs more material to work from.

What are the three topic types?

You decide how much of a topic is scripted, topic by topic. The same engine runs all three.

A creative topic gives the writer a principle and lets it find the words. A semi scripted topic holds your fixed lines in place and lets the engine fill around them. A structured topic scripts the hook, the beat and the discussion prompt, so the post comes out in a shape you already chose. If you have ever written a caption you were happy with and wanted more like it, structured is where that lives.

Will the caption fit the image, or come out cut off?

It fits, or it does not post. Every format declares its text limits, and a post that breaks them is parked for a person to look at instead of going out cut off.

That is the one place a human steps in, and it is a safety net rather than a review step. Everything that fits posts on schedule, and only the posts that break a limit wait for you.

What happens after it posts?

The post goes out through the route its platform page declares, and the run is logged. Today it publishes to pages and accounts you own on Facebook, Instagram, TikTok, X, YouTube, LinkedIn and Snapchat.

Every run writes a log with its status and how long it took, plus a log of every call it made to an outside service, and the admin console shows both. Every paid outside service has a daily cap, and a call over budget is refused and logged rather than quietly run up. Volume stays capped, timing is varied rather than landing on the same second every day, and one posting session runs per machine so two processes never fight over the same page. The reasoning behind those choices is on trust and account safety, and the platform by platform position is on platforms. If you want to see which kind of operator this fits, start at use cases, or read the straight answers on the FAQ.

What’s coming, and what is not built yet?

A full planner is designed, not built yet. None of this section runs today.

The design weighs every candidate for a posting time on a formula with plain terms: is there material that fits, is the subject moving right now, and how long has it been since it last ran. Rotation would stop the page posting the same frame twice in a row, and floors would keep weaker ideas in play so the page never narrows to one. A fourth term, how a post actually performed, is part of the design too, but no engagement data is collected.

Alongside the planner sits a decision log: for every post, which story won, what it beat and why. That is designed, not built yet either. When any of it ships, the changelog says so with a date, and this section moves up into what runs today.