Facebook · Comment · Oct 12, 2020
“If you come near my sister again I will put you in the hospital.”
REVIEWThreats
A direct threat of physical harm aimed at a real person.
Prune: Delete the comment.
Run a free social media background check on yourself. Find the forgotten like or comment before someone else does.
PruneSocial checks the posts, comments and likes in your own account exports against published screening categories, then helps you decide what to remove, change or keep.
Free · No account connections or passwords · No card
No exports yet? Request them now
Your ZIP opens in this browser. One preview per email. US residents 18+. Likes are scored on X and Instagram.
Sample report summary
Highest detected risk
Categories checked
ThreatsReview
ViolenceConsider
Trolling / HarassmentConsider
DrugsConsider
CrimeConsider
IntoleranceClear
SexClear
- Private self-review only
- No account connections
- Not shared with employers
- Free, then $49
For job seekers, working professionals and people in public roles.
Employers look. Some look back seven years.
-
They look
3 in 4
of hiring managers say they check applicants’ social media.
ResumeBuilder survey of 1,013 hiring managers, 2023 -
It costs offers
57%
of employers who researched candidates online found something that made them decide not to hire.
CareerBuilder and Harris Poll, 2018 -
They dig back
7 years
is the lookback screening vendors such as Checkr and Fama describe.
Vendor descriptions, with sources -
But vanishing backfires
47%
of employers are less likely to interview a candidate they can’t find online.
CareerBuilder and Harris Poll, 2018
Deleting everything isn’t the answer. Pruning is. In 127 news cases we collected (a curated set, not a study or a rate), 113 ended with the person losing the job, seat, admission or license. Some began with a like or a share, not something they wrote. See the cases.
Found Later · 127 news cases
It has happened in every kind of job.
Real people have lost jobs, seats and offers over an old post, like or comment. This is a curated set of reported cases across 19 kinds of work, not a study or a measure of anyone’s risk.
113 of 127 cases ended with the person losing the job, seat, admission or license.
Browse all 127 casesMost-scrutinized fields
- School boards and local office15
- Athletes10
- Teachers9
- Police9
- Media and entertainment9
Each case links to its news source. Names are left out on purpose.
Three steps
01
Request your export
Download your own data from each platform. Instagram and Facebook take hours; others a day or more.
02
Upload and get scored
Drop in your ZIPs. Each item is rated Review, Consider or Clear. Start with a free 50-item preview.
03
Prune each flag
Each flag gets one action, like removing a like or deleting a comment, with steps for that platform.
What we scan, and what we look for
Seven categories checked
- Crime
- Violence
- Threats
- Trolling / Harassment
- Intolerance
- 18+ sexual content
- Drugs
Movie quotes, sports talk, political opinions and recovery content aren’t screening flags. Politics, news events and workplace complaints are listed separately as reputation flags. See exactly what is.
Where it looks
| Platform | Posts | Comments | Likes | Bio |
|---|---|---|---|---|
| X (Twitter) | Scored | Scored | Scored | Scored |
| Scored | Scored | Captioned | Scored | |
| Scored | Scored | Names only | Not in export | |
| Scored | Scored | Not in export | Scored | |
| Scored | Scored | Not in export | Not in export | |
| Threads | Scored | Scored | Not in export | Not in export |
| YouTube | Scored | Scored | Not in export | Not in export |
| TikTok | Scored | Scored | Not in export | Not in export |
| Scored | Not in export | Not in export | Scored |
✓ scored · Captioned: only liked posts that have a caption · Names only: the Page or person, not the post · – not in that platform’s export. Posts includes captions, video titles and descriptions, and pins. See exactly what’s scored.
What a report shows
Illustrative examples, not a real person’s history.
Instagram · Liked post · Apr 20, 2022
“Asian drivers, am I right? Laughed for 10 minutes at this.”
CONSIDERIntolerance
A liked post that jokes about a stereotype of a whole group. Likes are part of your history.
Prune: Remove your like.
Facebook · Comment · Jun 17, 2025
“I’m voting against this tax increase and I think the mayor is wrong on this one.”
CLEARNo category match
A political opinion without insults or threats isn’t a screening category. It can still be listed separately as a reputation flag.
Your report arrives as a private PDF
Flags ranked strongest first, with one action each.
See a full sample report (PDF)Simple pricing
One payment. No subscription, no account.
30-day guarantee. Not useful? Email hello@prunesocial.com within 30 days of your first report (or 60 days from purchase if you haven’t run one) for a full refund, including any add-on. US residents 18 and older.
Text only, on purpose. It reads your words, not what photos or videos show, and doesn’t predict any employer’s decision. What’s scored.
Upload an export. Don’t hand over a login.
No password, no account connection, nothing to revoke later. Your ZIP opens in your browser and only text is sent.
How your data moves
- On your computer
Your ZIP is read in your browser. Only the text to be scored goes further.
- Scored
Anthropic’s commercial API rates that text. Photos, messages and passwords are never sent. API content isn’t used for training by default.
- Delivered, then deleted
Your PDF waits behind a private link, then is deleted after 14 days.
Anthropic normally deletes scored text within 30 days, with exceptions. Anyone with your full report link can open it. Full details.
The only wait is on the platforms. Reddit can take up to 30 days, so request it first, and start with whichever ZIP arrives first.
Frequently asked questions
Can't I just Google myself for free?
Googling shows what is public. It doesn't apply a screening-style rubric or review your likes and comments. PruneSocial scores the text in your own exports and explains each flag. For the DIY route, see our free self-check guide.
How do I get my content to you?
Request a copy of your data from each platform; the export guides show how. Drop the ZIP on the scan page. It opens on your device. Only the text to be scored (posts, comments, likes and bios, with each item’s date, type, and a link to it when the export includes one) is sent. Photos, messages and passwords are not.
Can my AI agent do this for me?
Yes, if it can work in your own browser while you’re signed in. It can request your exports, add them to the scan page after you tick the consent boxes, and work through your report’s fixes using the agent cleanup file that comes with every paid report. It asks you before each change, and PruneSocial never signs in to your accounts. Prompts and steps.
How do you know what screeners flag?
Several screening vendors publicly describe the categories they flag and lookbacks of up to seven years. PruneSocial builds its own self-review rubric from those descriptions and doesn't reproduce any vendor's classifier. Methodology and sources.
My accounts are private. Do I still need this?
Often yes. Your name, photo and bio are usually public, comments on public pages are public, and settings change. Your export also includes private content, so a report can flag items the public can't see today. A flag means a match to a category, not that an employer can see it.
What happens to my data?
Your content is used to build your report, then deleted from our systems. The finished PDF is stored encrypted for 14 days behind a private link. Scoring runs through Anthropic's commercial API, which doesn't train on API content by default, subject to its retention and safety exceptions. Privacy Policy.
Some fields review online history more closely
Know what’s in your history before you apply
Start with a free 50-item preview.