The post that ends your campaign is already up.
Most brands find it after launch, in the comments. CreatorScore reads every post, every spoken word and every frame a creator has published, on every platform, and shows you the problem before you sign.
Every post, spoken word and frame, on every platform the creator is on.
Your team read the last thirty posts. She has 2,481. The one that matters is a voiceover from March 2024, and the caption over it says “my morning tea ritual”.
▸ What not checking costs
You find out on day 14. Or on day −2.
Old posts don’t stay old. A clip from two years ago gets found the week your logo is next to it, and the customers who find it tag you, not the creator. The post was public the whole time.
- The full history on every platform, not the last month
- Each finding names the post, the line and the date
- $9.99 a creator, less in volume
▸ Every layer of every post
The caption is clean. The audio isn’t.
A person scrolling reads captions. The risk is usually in what was said out loud, the text on screen or a frame in the middle. We read each layer of each post separately: speech is transcribed, frames are checked by vision models, on-screen text is read.
- Hate, explicit content, violence, drugs, scams and misinformation
- Undisclosed paid posts, from the platform’s own partnership record
- Judged in context, so comedy isn’t read like a kids’ channel

- Caption“my morning tea ritual ☕”…
- Speech“throw out your meds, this tea is all you need”…
- On screen“no more pills 💊” text overlay…
- FramesKitchen, stove, oranges, a mug…
- Comments1,904 read · 3% bot-like…
▸ Some problems override the score
A great average can’t hide a serious problem.
Seven agents score a creator from 1 to 100. If a confirmed problem turns up, the score is capped however good the rest is. Each cap needs evidence that holds up: one ambiguous frame or a single classifier hit never caps anyone.
- Bot commenters, engagement pods, hate the creator produced
- Repeated explicit content, missing FTC disclosure
- Every cap traced to the posts behind it
▸ After you sign
The contract is signed. She posts again tonight.
A check is a snapshot. Monitoring reads what the creator posts after it, every morning, and emails your team when a critical or high finding holds up on review, with the post attached. Everything else goes in a Monday summary.
- New posts read daily, X twice a day
- Instagram stories saved, so early deletions still show
- Full rescore every 30 days. $19.98 a month per creator
- Mon 06:104 new posts readTikTok, Instagram. Nothing found.
- Tue 06:12Instagram story savedDeleted 3 hours after posting. We kept a copy.
- Wed 06:082 new posts readYouTube transcript checked. Nothing found.
- Thu 06:15Critical findingA slur in a TikTok video at 00:31, from the transcript, upheld on review.
Before the contract. After it.
A check takes a couple of minutes to order. The campaign launches whether you ran it or not.
▸ Two ways to check
A score to decide fast. A background check to go deep.
One score from 1 to 100 across every platform, with the post behind each finding. Keep it current for $9.99 a month, or add daily monitoring for $19.98.
Check a creator- Seven agents, each scored on its own
- Bot and engagement-pod checks on the comments
- $8.99 each from your 51st creator in a month, $7.99 from the 201st
A full read of everything the creator has posted, every finding listed by category with its evidence. No score, just what was found.
Run a background check- The whole history, not a sample
- Legal matters shown with how they ended
- Deleted posts recovered on the Verified tier
▸ Questions
What brand teams ask first.
What is influencer brand safety?
Influencer brand safety means checking a creator's content for risk before your brand is attached to it, and after. That covers their posts, what they say in videos, text on screen, images and comments, looking for hate speech, explicit content, misinformation, scams, controversy and fake engagement.
How does CreatorScore find unsafe content?
Language models read captions, comments and video transcripts. Vision models check images and video frames. Text on screen is read separately, because it often says something the caption doesn't. The Content Risk Agent combines nine signals: hate signals (18.4%), scams and deception (18.4%), explicit content (13.8%), flag severity (9.2%), web controversy (9.2%), visual risk (9.2%), misinformation (9.2%), what is said in videos (8%) and profanity (4.6%). If a signal can't be measured, its weight moves to the ones that were, so nobody loses points for missing data.
How long does a check take?
You order it in a couple of minutes. The time to read a creator depends on how much they have posted; the dashboard shows progress, and you get the score and every finding as soon as it's done.
What happens after we sign the creator?
With monitoring ($19.98 a month per creator), new posts are read every morning across the creator's platforms and X twice a day, and Instagram stories are saved so you can see one deleted early. A critical or high finding that holds up on review is emailed to your team with the post; everything else goes in a Monday summary. The full rescore, including comments, frame-by-frame video and bot checks, runs every 30 days. Live streams aren't monitored.
What is the difference between a CreatorScore and a background check?
A CreatorScore ($9.99) is a 1 to 100 score across every platform the creator is on, with the findings behind it, and it can be kept current every month. A Social Background Check ($29.99) is a one-time read of a creator's full history that lists what was found by category with the evidence. It doesn't produce a score.
Your next creator
posts again tonight.
Check them before the contract. Keep reading after it.