Elixir Profanity Filter

Req.post!(
  "https://api.theprofanityapi.com/v1/check",
  headers: [{"authorization", "Bearer #{System.get_env("PROFANITY_API_KEY")}"}],
  json: %{text: text, context: "chat", mode: "balanced"}
).body
# => %{"flagged" => false, "intent" => "safe", "score" => 0.02}

Why context-aware matters

A plain word list flags "this game is sick" or medical text full of anatomical terms. The Profanity API runs a five-layer pipeline that scores intent, so "die" in a gaming taunt and "die" in a real threat are handled differently. Pick a context and the engine adjusts strictness automatically.