What we run
X growth, run for you.
X rewards replies, not posts. Scheduled posting is explicitly permitted by X, and the API is pay per request. But AI-generated replies require prior approval from X, and replies are where the conversation happens. So an operator writes and sends them from your account.
Last reviewed 27 August 2026
The failure mode
What breaks when AI does this alone.
Posting is the permitted part. X developer guidelines explicitly list an automated account posting scheduled content as good-faith automation, and the API is priced per request rather than per subscription. What breaks is the half that actually grows an account. X states that deploying AI-generated replies without prior approval is a violation even when the content is helpful, and it separately prohibits posting identical content across multiple accounts and auto-replying to keyword mentions. Replies are the growth mechanism and they are the gated surface.
What our operators do here
- Write and schedule original posts in the founder voice, approved before they ship
- Reply by hand in conversations where the founder has something real to add
- Run quote-posts and threads that take an actual position, not a summary
- Build and maintain a list of accounts worth being visible to
- Handle DMs and route qualified ones to a call
- Decline pile-on threads and anything where a reply would age badly
Volume
What ships in a typical week.
| What | Typical volume |
|---|---|
| Original posts and threads | 5 to 10 |
| Hand-written replies in other conversations | 20 to 40 |
| Conversations monitored for a real opening | Ongoing, daily |
| Threads declined as a bad fit | Logged and reported |
Volumes are typical rather than contractual. They move with what the channel is actually returning.
What good looks like
The metric that matters
Replies from accounts your buyers already follow, and profile visits from them
Realistic timeline
8 to 12 weeks. Follower count moves earlier and means less than it looks.
The gated surface is the one that works
X is unusual among the channels we run, because the automation rules and the growth mechanics point in opposite directions.
X's developer guidelines are permissive about posting. They explicitly list an automated account posting scheduled content as an example of good-faith automation, alongside an app posting RSS feed updates on behalf of a user. Nothing about scheduling drafts is against the rules.
The same guidelines then say that deploying AI-generated replies without prior approval from X is a violation, and add that this holds even if the content itself is helpful. That is the sentence that decides how this channel has to be run, because replies are how accounts actually grow on X.
Why replies and not posts
A post from an account with four hundred followers reaches roughly four hundred people, most of whom are asleep. A good reply in a thread that already has traction reaches everyone reading that thread, which is a different order of magnitude and a much better-qualified audience.
That is not a growth hack, it is just how the surface works. Which is why every serious X account you can think of was built in other people's replies before it was built in its own timeline.
So the channel that is most tempting to automate is the one where automation is explicitly approval-gated, and the part you are allowed to automate is the part that does the least.
What is allowed, precisely
| Activity | X position |
|---|---|
| Automated account posts scheduled content | Explicitly permitted as good-faith automation |
| App posts RSS updates for a user | Explicitly permitted |
| AI-generated replies | Requires prior approval from X |
| Identical content across multiple accounts | Prohibited |
| Auto-replies to keyword mentions | Prohibited |
| Auto-DMs to new followers | Prohibited |
Worth noting for anyone budgeting: X now prices the API per request rather than by tier. As retrieved on 27 August 2026, a post costs 0.015 dollars per request and a post containing a URL costs 0.200 dollars, about thirteen times more. X states prices may change. That pricing gap is a reasonable proxy for how the platform feels about outbound links.
How we run it
The AI layer handles research and drafting. It watches the accounts and topics that matter to you, surfaces conversations where you have a genuine position, and drafts posts and threads against your voice profile.
An operator does the rest by hand. They decide which conversations are worth entering, which is mostly a filtering job. They write the reply themselves, in your voice, from your account. They stay in the thread when someone pushes back, because a reply that provokes a good objection and then goes quiet is worse than silence.
You approve the original posts. Replies move too fast for an approval loop, so we work from a standing brief of what you will and will not claim, and we escalate anything outside it.
What we will not do
We will not auto-reply to keyword mentions, auto-DM your new followers, run duplicate content across accounts, or post into unrelated trending topics for visibility. X prohibits all four by name, and they are also the behaviours that make an account read as a bot to humans long before the platform notices.
We will not pile onto a public argument for reach. It works, briefly, and then it is attached to your company name for as long as the post exists.
And we will not report follower growth as the headline. We report who replied.
X questions
Does X allow AI to run a startup account?
What does the X API actually cost now?
Can you run the same content on X and LinkedIn?
Do you buy followers or run engagement pods?
How long does X take to matter?
See what we would run on X.
Thirty minutes. You leave with the thirty-day plan either way.
30 minutes. You leave with the plan either way.