NSFW Text Detection

Detect sexually explicit language, hate speech, harassment, self-harm, and other harmful categories in user-generated text.

POST /nsfw-text/detect

Authentication

Bearer token in the Authorization header.

Request

Content-Type: application/json

Field	Type	Required	Description
text	string	Required	The text to classify. Long text is automatically chunked.

Response

Field	Type	Description
score	number	Overall NSFW probability (0.0 – 1.0).
is_nsfw	boolean	Whether the text is classified as NSFW.
categories	object	Per-category scores, e.g. {"sexual": 0.01, "hate": 0.02, "violence": 0.0, "harassment": 0.0, "self_harm": 0.0}.
verdict	string	Human-readable verdict summarising the result.

Example

Request

curl -X POST https://api.cacuda.com/nsfw-text/detect \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"text": "Hello, how are you today?"}'

Response 200 OK

{
  "score": 0.01,
  "is_nsfw": false,
  "categories": {
    "sexual": 0.01,
    "hate": 0,
    "violence": 0,
    "harassment": 0,
    "self_harm": 0
  },
  "verdict": "safe"
}

Error Codes

Code	Description
400	Missing or empty text field.
401	Missing or invalid API key.
402	Insufficient credits.
503	Detection service unavailable.

Notes

Text longer than ~450 tokens is automatically split into chunks. Each chunk consumes one credit.

Category keys may vary by model ensemble version.

Try it in the scanner

Test this endpoint interactively with real data. No code required.