AI Content Moderation

Layered screening on every post and comment, tuned for sport so competitive talk is not mistaken for abuse.

Status

Live in production

Covers

Every post and comment, text and video

Tuned for

Sport-specific context

Built by

CTO-led, with our five-person ANU engineering team

Text and video screening

Sport-tuned thresholds

What it does

Every post and comment passes through layered screening, text and video both. Off-the-shelf filters fail at sport in both directions: they flag ordinary football language as violence and miss genuine abuse dressed up as banter. Neylo’s thresholds are tuned on sport content specifically, so the system reads context instead of matching keywords.

For athletes

Comment sections on sports content are famously rough, and the people on the receiving end here are teenagers. Screening runs before harm reaches the athlete. A kid should be able to post a missed tackle without copping abuse for it, and on Neylo they can.

For scouts and clubs

A club cannot put its staff on a platform where minors get abused. Full stop. Moderation is the difference between a tool a club can officially adopt and one it has to pretend not to use. It also protects the scouting signal itself, because comment sections full of noise bury the football.

Where it fits

Moderation is the public face of the same principle as our trust and safety layer: growth is worthless if the platform is not safe for the people on it. It is also why moderation shipped before the AI features that will drive growth did.

How it’s built

Layered rather than single-pass. Fast screening runs on everything, and deeper review runs on whatever gets flagged. The flagged-content screens in our product shots are not mockups. This is live in production on every post and comment today, with thresholds maintained by the engineering team as the content base grows.