The Profanity API is a text moderation service that tells banter from abuse. A single POST request returns a verdict, a score, the intent and the layer that made the call. Messages pass through a blocklist, a word list, word and phrase similarity, and a judgment layer, and an LLM runs only when text is still ambiguous. It normalizes leetspeak, dots and look-alike letters, and it knows about 50,000+ harmful phrases and 1,500+ profane terms. A free plan covers 300 requests a month, and Pro is pay as you go at $0.002 per request.
It is built for developers and product teams who run places where people type to each other: game studios with lobby chat, community and forum owners, social app builders, support teams and anyone shipping a kids' app. Engineers can integrate it from any language that makes HTTP requests, with code samples in four languages, so a small team can add moderation in an afternoon. Trust and safety staff can use the layer details and raw JSON to understand why a message was flagged.
A gaming chat can allow trash talk such as a joke about destroying a team while still flagging a veiled threat like knowing where someone lives. A kids' app can use the child_safe context for stricter rules, and a professional workspace can apply different thresholds. Live chat can use instant mode, which runs in under a millisecond, while comment sections use smart mode for fewer false bans. Operators can block a message, warn the author or send it to a human reviewer, and keep a custom allowlist for place names such as Scunthorpe.
