Ittybit

Overview

The NSFW Task detects explicit or unsafe content in images and videos.

It analyzes the visual content and returns confidence scores for several categories such as nudity, sexual content, violence, and gore.

When a task completes, it creates an Intelligence file with kind: "nsfw" and a .json output containing the detection results.

Example Output

{
  "id": "file_qrstuvwx9012",
  "object": "intelligence",
  "kind": "nsfw",
  "detected": true,
  "nudity": 0.82,
  "sexual": 0.24,
  "violence": 0.64,
  "gore": 0.04,
  "confidence": 0.82,
  "created": "2025-01-01T01:23:45Z",
  "updated": "2025-01-01T01:23:45Z"
}

Each field represents the confidence level (0–1) for that specific category.

A higher value means stronger detection likelihood.

Creating an NSFW Task

You can create an NSFW detection task for any image or video file using the ittybit SDK or a direct API request.

import { IttybitClient } from "@ittybit/sdk";

const ittybit = new IttybitClient({
  apiKey: process.env.ITTYBIT_API_KEY!
});

const task = await ittybit.tasks.create({
  kind: "nsfw",
  url: "https://example.com/image-or-video.mp4",
  description: "NSFW content detection analysis",
  webhook_url: "https://your-app.com/nsfw-webhook"
});

console.log("Task created:", task.id);
console.log("Status:", task.status);

Webhook Example

When the task completes, ittybit will send a POST request to your webhook_url with the results.

You can use this to automatically flag, moderate, or remove content.

app.post("/nsfw-webhook", async (req, res) => {
  const { id, kind, status, results } = req.body || {};

  if (kind !== "nsfw" || status !== "completed") {
    return res.status(200).send("Not a completed NSFW task");
  }

  // Example: update Supabase database
  const { error } = await supabase
    .from("uploads")
    .update({
      nsfw_detected: results?.is_nsfw || false,
      nsfw_confidence: results?.confidence || 0,
      processed_at: new Date().toISOString()
    })
    .eq("task_id", id);

  if (error) {
    console.error(error);
    return res.status(500).send("Update failed");
  }

  res.status(200).send("NSFW status updated in DB");
});

This example mirrors the production-ready implementation from

Check every Supabase upload for NSFW content.

File Structure

Property	Type	Description
id	string	Unique file ID for the Intelligence file.
object	string	Always "intelligence".
kind	string	Always "nsfw".
detected	boolean	Whether any unsafe content was detected.
nudity	number	Confidence score (0–1) for nudity detection.
sexual	number	Confidence score (0–1) for sexual activity or context.
violence	number	Confidence score (0–1) for violent content.
gore	number	Confidence score (0–1) for gore or graphic imagery.
confidence	number	Overall confidence score for the detection result.
created / updated	string (ISO 8601)	Timestamps for creation and last update.

Supported Inputs

NSFW tasks work with both image and video sources:

Image: .jpg, .jpeg, .png, .webp
Video: .mp4, .mov, .webm

Common Use Cases

User-generated content moderation
Automatic content filtering before publishing
Flagging or blurring unsafe media
Age-restricted platform compliance

Example Workflow Automation

You can combine NSFW detection with an automation workflow to process all new uploads automatically:

{
  "name": "Moderate new uploads",
  "trigger": {
    "kind": "event",
    "event": "media.created"
  },
  "workflow": [
    { "kind": "nsfw", "ref": "safety-check" }
  ],
  "status": "active"
}

This automation will run an NSFW task on every newly created media file.

NSFW

On this page