How it works

modwall is a single endpoint: you send an image and get back a score per category: NSFW (nudity / adult content) always, and weapons and violence/gore (beta) — when you enable them in your profile. We do not detect offensive gestures. Text is scored by a separate model for toxicity. We process the image in memory and store it nowhere — we log only metadata (time, score, size).

Three steps

1. Create an account and generate an API key in the panel.
2. Send the image to POST /v1/moderate with the header Authorization: Bearer sk_live_….
3. You receive raw scores (nsfw/safe) and a decision computed according to your policy.

You set the threshold and mode on your side — a kids' app stricter than an adults' forum. The monitor mode lets you plug modwall in with no risk: it blocks nothing, it only shows what it would flag. We always return the raw result.