What Counts as Original Content on X and How Do You Stay Eligible?
The originality rule decides which of your posts earn, and most of the disqualifiers have nothing to do with copying.
Original content on X means material whose primary value comes from your own work rather than from someone else's. That single sentence is the whole rule, and it is the sentence that decides whether a post feeds your qualified impressions or sits there earning nothing. X wrote the originality requirement into the Original Content Rewards terms that took effect on 7 August 2026 and announced the change publicly the next day. Under the programme that came before it, no such requirement existed at all.
The practical consequence is uneven. If you write your own posts and shoot your own photos, nothing about your workflow changes. If you run an aggregator feed, a quote-post account or a pipeline that moves clips from another platform onto X, the rule reaches directly into your format. And there is a second group of disqualifiers, unrelated to copying, that catches accounts which have never reposted anything in their lives.
What Does X Actually Count as Original Content?
X gives a positive list before it gives a prohibition, which is the more useful place to start. Original content covers text you wrote yourself, including original reporting and analysis. It covers photographs and video you shot. Graphics, illustrations and memes you designed count as well. And the list explicitly reaches commentary and reactions that carry a genuine contribution, with X stating that commentary sits at the heart of the platform.
That last item is the one people misread. Reacting to something you did not make is not automatically disqualified, and X went out of its way to say so. A post can be built entirely on a news event, a video someone else published or a cultural moment nobody owns, and still count, as long as what you add is the thing carrying the value. Think of a sports account that posts a screenshot of a substitution and writes four hundred words on why the manager made it. The screenshot is not the product. The read is.
Where accounts get caught is the gap between adding something and adding something that matters. X has not disclosed whether originality assessment runs through an automated model, a human reviewer or both, so there is no published scoring rubric to game. The workable approach is to ask what a reader takes from your post that they could not take from the thing you built it on. If the honest answer is nothing, the post is doing no work for you.
The Four Categories That Stop a Post From Earning
X named four categories of content that do not generate qualified impressions. They overlap in practice, and a single post can land in two of them at once.
| Category | What it covers | Concrete example |
|---|---|---|
| Copied content | Someone else's text or image published as is | Downloading a clip and reuploading it to your account |
| Minimally modified | Surface changes to material you did not make | Swapping a word or two, adding a filter, changing video speed, dropping a text layer on top |
| Aggregated content | Other people's work assembled without a new angle | A daily roundup feed that lists items and adds no reading of them |
| Cross-platform reposts | Material lifted from another platform onto X | Pulling a video off a short-form app and posting it as your own |
The governing rule X wrote sits underneath all four. If the primary value of a post comes from someone else's work and little was done to transform it, the post does not produce qualified impressions. Notice what the rule does not say. Nowhere does it forbid you from using other people's material, and nowhere does it set a percentage. What the rule sets is a test about where the value lives, which is why two posts built on the same clip can land on opposite sides of it.
X also supplied an example of what falls short, and it is worth sitting with because it is the most common workaround people reach for. Adding a caption or a text overlay that only describes what is already happening on screen is not enough. Writing "he did not see that coming" over a clip of someone not seeing something coming adds a label, not a contribution. The same logic applies to a repost with a one-word reaction on it.
What Kind of Transformation Does X Accept?
X did not leave transformation as an abstraction. It named four things that qualify, and they map cleanly onto formats that already exist on the platform. You can paraphrase a news story. You can share your own take on a cultural moment. You can add context that other people do not have. And you can meaningfully transform content through humour or creative fiction.
Read those four together and a pattern shows up. Each one requires you to bring in something that was not in the source material, whether that is knowledge, judgement, a point of view or an invented frame. Paraphrasing a news story is not copying a headline, it is restating the substance in your own words. Adding context is the strongest of the four for accounts built around other people's material, because context is the one input that scales with what you actually know.
A worked scenario helps. Suppose a company announces a product change and the announcement is public. An account that screenshots the announcement and writes "big if true" has transformed nothing. An account that screenshots the same announcement and explains what the change breaks for existing users, drawing on the previous version nobody else remembers, has added context others do not have. Same source, same screenshot, completely different position under the rule. The second account is producing original commentary and the first is producing a caption.
How Should Aggregator and Quote Accounts Adapt?
Aggregator accounts have the hardest adjustment, and X has already shown its hand on them. In the payout cycle around 11 and 12 April 2026, aggregator accounts saw their payouts reduced to 60 percent of the usual level, with a further 20 percent reduction announced for the cycle after that. Accounts that stamped "BREAKING" onto every post were singled out in the same action. That happened before the originality rule was even written into the terms.
The adaptation is not a rebrand, it is a change to what each post contains. Three moves cover most cases. First, stop posting the source item alone and start posting your reading of it, so the analysis carries the post and the source supports it. Second, credit the original creator, which does not by itself make a post original but does keep you out of the copied content category and out of trouble that has nothing to do with monetisation. Third, cut the reflex posts. A feed that publishes forty items a day cannot add a genuine contribution to forty items a day, and the ones that carry nothing are not earning anyway.
Picture a tech news feed that currently posts thirty items a day, each one a screenshot and a headline. Under the four categories, almost all thirty land in aggregated content, because nothing is added beyond assembly. The same account posting six items a day, each with two or three sentences explaining why the item matters to a specific group of readers, is running a different product with the same source material. Fewer posts, more of them capable of earning.
Quote-post accounts have an easier path, because the quote format already leaves room for the thing X is asking for. The failure mode is habitual brevity. If your quote posts are mostly a laughing emoji, a single word or a restatement of what the quoted post said, you are running a minimally modified operation with extra steps. Treat the quote as a prompt for your own argument and the format works in your favour, since you get attribution and your own commentary in the same unit. Cross-platform pipelines are the one habit with no rescue. Material lifted from another platform is named as a disqualifying category, and no caption fixes it.
The Disqualifiers That Rarely Make the Guides
Copying is the disqualifier everyone writes about. The rest of the list catches accounts that produce every word themselves, and it is where genuine creators lose payouts without understanding why.
| Disqualifier | What it looks like in practice |
|---|---|
| Monetisation coaching content | Posts focused only on making money on the platform or maximising payouts receive no payment |
| Helpful Community Note | A post that has received a "helpful" Community Note receives no payment |
| Engagement bait | Repeatedly asking for likes, replies, bookmarks, follows or reposts, treated as an ongoing violation |
| Misleading information | Posts carrying information that misleads readers |
| Automated means | Content created or posted by automated tools receives no payment |
The monetisation coaching exclusion deserves a moment, because it removes an entire cottage industry from the payout pool. Content whose subject is how to earn on the platform, or how to push your own payouts higher, does not get paid for doing it. You can still publish that material. It simply stops being a revenue source, which changes the economics of the whole genre.
The Community Note rule is the one most likely to surprise a working creator. You can write something entirely yourself, break nothing, copy nobody, and lose the payout on that post because readers attached a note the system rated as helpful. Accuracy is now a monetisation input rather than a reputational one. Engagement bait works on similar logic, since the "follow for part two" pattern is repeatable at scale and X treats repetition of it as a continuing violation rather than a single mistake.
There is also a narrower rule with a fixed start date. From 3 March 2026, creators who share AI generated video about armed conflict without labelling it are removed from the programme for 90 days, with permanent removal for a repeat. That rule is specific to unlabelled synthetic footage of conflict, and it does not settle the broader question.
Which brings up what X has not said. The list includes content created or posted by automated means, and that wording could plausibly reach AI generated posts, but X did not name artificial intelligence anywhere in the originality rules. No official statement declares AI assisted writing or AI generated imagery either eligible or ineligible. Treat it as unresolved rather than as a settled permission, keep a record of your own input into anything you publish, and label synthetic media where a rule already requires it.
What Happens If X Decides Your Views Were Inflated?
The enforcement clause is short and it is the reason the originality rule has teeth. A creator found to have artificially inflated their views forfeits every amount they had earned and is removed from the programme. X also reserves the right to claw back and offset past payouts, which means money already paid is not permanently yours if the account is later found to have gamed the counter.
The offset mechanism is the part worth planning around. Forfeiture only touches the balance you have not received yet, so its worst case is a quarter of lost income. A clawback reaches backwards into cycles you already spent, and an offset can take it out of future earnings rather than asking you to write a cheque. An account that ran an inflated pattern for months and then cleaned up its act is not automatically clear of it.
Nothing in that clause should worry a creator who writes their own posts and lets distribution happen on its own. Where it does bite is any arrangement that pushes numbers artificially, and the practical instruction is to keep your growth attributable to things you can point at. If a spike in impressions has an explanation you would be comfortable stating out loud, the enforcement clause is not about you.
How X, YouTube and TikTok Handle Originality
X is not moving alone here, and seeing the three programmes side by side makes the direction clearer than any single rule does.
| Platform | Entry requirement | What the payment attaches to |
|---|---|---|
| X Original Content Rewards | 500 verified followers, 500,000 verified Home Timeline impressions in 90 days, a paid Premium subscription | Qualified impressions, with four named disqualifying categories |
| YouTube Partner Programme | 1,000 subscribers plus either 4,000 watch hours in 365 days or 10 million Shorts views in 90 days, rising to 8,000 hours or 20 million views from 1 February 2027 | Advertising revenue share |
| TikTok Creator Rewards | 10,000 followers, 100,000 video views in the last 30 days, 18 or over, limited country list | Qualified views, with video required to run at least one minute and be original |
Three threads run through all of it. Every one of the three programmes has written an originality requirement into its rules rather than leaving it to enforcement discretion. Thresholds are being raised across 2026, with YouTube's next increase already dated. And payment in each case attaches to a filtered count rather than a raw view number, so the metric that appears on your post is not the metric that pays you.
For a creator working across platforms, the useful conclusion is that a single body of genuinely original work now travels further than a repost operation running on three feeds at once. TikTok makes originality an explicit condition alongside a length floor. X makes it a condition with four named exclusions. The formats that struggled hardest under X's April payout reductions are the same formats these rules were written to price out, and the ones that hold up are the accounts where the person behind them is visibly doing something.
This article was last updated on 20 August 2026 thursday. Today, 61 visitors read this article.
