▸ For legal and compliance review
What each output is and is not, how coverage is stated, what every finding carries, which rules the regulated-industry codes consider, and how to verify the exact report you were handed.
▸ The four outputs
Two products share one pipeline. A background check never produces a score; a score never reads a background check.
A one-time inspection of a creator's public content history, categorised into brand-safety findings with the post, frame or transcript behind each one. A public-record mode reads press and legal coverage of a public figure and states every matter with its outcome.
What it is not
It produces no score and makes no recommendation. It is not a consumer report and is not FCRA-compliant.
A 1 to 100 trust score across a creator's whole footprint, from seven weighted agents, with every flag tied to the post it came from and a two-sentence narrative. Monitoring refreshes it on a schedule.
What it is not
It is not a compliance determination. A high score says the footprint showed no conduct we detect as risk; it does not say a campaign is lawful.
A creator checked clause by clause against a code: the FTC Endorsement Guides, a brand-safety baseline, an alcohol baseline, a customer's own code, or one of four regulated-industry codes. Each clause shows a flag with its evidence, a pass over a stated scope, or not assessed.
What it is not
A clause names the rule it considers; it never states that a creator is compliant or in breach of it.
What a creator's audience says in the comments: sentiment over time, how sponsored posts land, themes, brands the audience names, repeated issues and recurring topics, each with the comments behind it.
What it is not
It is read from stored comments and a model's grouping of them; it is context for a reviewer, not a finding against the creator.
▸ Coverage
Every governance clause carries one of four coverage grades, derived from which detectors are live, never typed by hand.
Every detector behind the clause is live and nothing in the clause's scope is left unexamined. A pass is a pass over the whole clause.
A verdict is possible, but part of the clause is not examined. The report states the gap under the pass as 'Scope of this pass'.
A detector emits findings but cannot ground a pass. A finding is shown; silence is never a pass.
No signal reaches a verdict. The clause says why, and what a reviewer does instead. Not assessed is never a pass.
▸ Every finding
A count is never a finding. If we cannot show the proof, we do not apply the penalty.
▸ Regulated industries
Each is CreatorScore's plain-language reading of the rules a reviewer in that sector checks a creator against. A clause shows the rule it considers under its requirement and states no legal conclusion. Clauses we have no signal for yet, such as fair balance, licensing and audience age, read not assessed and say what a reviewer does instead.
FDA 21 CFR 202.1 and the 2025 direct-to-consumer letters, FTC Act section 5 and 16 CFR Part 255, Australia's TGA Code section 24, UK CAP Code 12
Securities Act section 17(b), FINRA Rule 2210, SEC Marketing Rule 206(4)-1, FTC Part 255, CFPB UDAAP, UK FSMA section 21
Securities Act section 17(b), FTC Parts 255 and 465, UK FSMA section 21 as applied to cryptoassets since October 2023
state producer-licensing and advertising rules on the NAIC models, CMS Medicare marketing rules at 42 CFR 422.2260, FTC Part 255
▸ Limitations
▸ Provenance, verification and disputes
Write to [email protected] with the report ID and we answer with the record behind it.
CreatorScore