Service 2 · AI review
Check every old tweet against X's own rules
The AI review reads every post and reply you have made against X's published rules, shows each risky one with the rule and the reason, and deletes only the ones you approve.
In short
PrunePosts' AI review checks each of your X posts, replies and quote posts against the full text of X's 18 published rule policies. A fast pass reads every post, a second pass re-checks anything flagged, and each flag names the rule, a severity and a reason. You keep or delete each one. Try 500 posts free; full passes start at $19 for 3,200 posts.
What it checks
The full text of X's published rules, taken from X's Help Center: 18 policies in four groups.
Safety
- Violent speech
- Violent & hateful entities
- Child sexual exploitation
- Abuse & harassment
- Hateful conduct
- Perpetrators of violent attacks
- Suicide & self-harm
- Adult & sensitive media
- Illegal or regulated goods
Privacy
- Private information
- Non-consensual nudity
Authenticity
- Platform manipulation & spam
- Civic integrity
- Misleading & deceptive identities
- Synthetic & manipulated media
- Financial scams
Intellectual property
- Copyright
- Trademark
How the review works
Load your posts
Sign in with X for your newest posts, and upload your free X archive for the rest of your history.
A fast pass reads every post
Each post, reply and quote post is checked against the full text of X's published rules. Reposts are skipped by default because they are other people's words.
A second look at anything flagged
Flagged posts get a second, slower check before you see them, which weeds out false alarms.
You review each flag
Every flagged post shows the rule it may break, a severity, a confidence and a one-line reason. Choose Keep or Delete.
Practice, then confirm
Run a practice round, download a backup, then confirm. Only the posts you approved are deleted.
What a flag looks like
@example_fan say that one more time and I'll find you after the game and break your face
Reads as a threat of physical harm, which X's violent speech policy prohibits.
What it won't do
- Promise you won't be suspended. X decides, and deleting a post doesn't undo enforcement that already happened.
- Read images or links. Posts that depend on them are marked Possible and never pre-selected.
- Judge behavior. Suspensions for spam or inauthentic behavior are about what an account does, not what it said.
- Delete on its own. Every deletion is one you approved.
Where your posts go
- Your archive is opened in your browser; only the posts files are uploaded, never your messages, photos or videos.
- Post text is sent to an AI model only to be checked. PrunePosts never uses your posts to train AI.
- X receives only the delete requests you approve.
- Your X sign-in is stored encrypted. When a post is deleted, PrunePosts erases its copy, and signing out can erase all your data.
What it costs
Prices in US$. There is no free tier: every post read from X, every deletion and every AI check costs money, so each is paid for out of your plan, and work pauses at the plan's limit instead of charging you more.
| Plan | Price | Billing | What you get |
|---|---|---|---|
| Starter | $4.99 a month or $49.99 a year | Subscription; renews until you cancel | About 340 deletions or 1,700 AI checks every month |
| Plus | $9.99 a month or $99.99 a year | Subscription; renews until you cancel | About 780 deletions or 3,900 AI checks every month |
| Pro | $14.99 a month or $149.99 a year | Subscription; renews until you cancel | About 1,200 deletions or 6,000 AI checks every month |
| Max | $29.99 a month or $299.99 a year | Subscription; renews until you cancel | About 2,500 deletions or 12,000 AI checks every month |
| Clean-up Pass | $49 | One-time; 30 days, never renews | About 4,800 deletions or 24,000 AI checks in 30 days |
| Clean-up Pass XL | $149 | One-time; 30 days, never renews | About 15,000 deletions or 75,000 AI checks in 30 days |
Uploading your X archive costs nothing, while reading posts through X's API uses your plan at X's price of half a cent a post. What uses your plan, and what happens at the limit.
Coming soon Finding posts by your own topics (for example, everything that promoted a company), and posts that got ratioed.
Questions
Will my old tweets get me suspended?
No tool can tell you for sure, because X makes that decision. What you can do is find the posts most likely to break X's rules before someone reports them. X says it suspends accounts for repeated violations or for one severe violation, so removing rule-breaking posts leaves less to count against you.
How do I find bad tweets from years ago?
Search X for from:yourhandle plus a word you regret, or download your X archive and search it. To check everything at once, the AI review reads every post in your archive against X's rules and lists only the ones that may break one.
Does the AI decide what gets deleted?
No. It only flags posts and explains why. You choose Keep or Delete for each one, posts that depend on context it can't see are never pre-selected, and nothing is deleted until you confirm.
What can't the AI review catch?
It reads text, not images, videos or linked pages, and it can't see behavior such as spammy following, which X also enforces. If the AI provider's own safety filter refuses to read a post, PrunePosts says so and leaves that post for you to check yourself.
Is this an AI writing detector?
No. It doesn't guess whether a post was written by AI. It checks your own posts against the rules X publishes for everyone on X.
Is PrunePosts approved by X?
No. PrunePosts is not affiliated with X Corp. and is not endorsed by X. It reads X's public rules and deletes through X's official API.
Sources
Checked October 22, 2026.