AI Image Safety Checker

Detect image safety and NSFW risk locally in the browser with an open-source ONNX model, preferring WebNN when supported while never uploading images

Category

Image Tools

How to Use

  1. Drag or select a PNG, JPG, or WebP image
  2. Choose Auto backend to try WebNN first with WASM fallback
  3. Click Check Image Safety to view neutral, porn, sexy, and related probabilities
  4. Use the risk score as a pre-screening signal and review important results manually

Examples

  • Pre-check before publishing

    Input: Upload a user avatar or community image

    Output: Returns NSFW-related probabilities and an aggregate risk score

  • Private local review

    Input: You do not want to send a sensitive image to a third-party service

    Output: Runs the check in the browser without uploading the image

FAQ

Is the checked image uploaded?
No. The image is decoded and classified locally in the browser; model weights are downloaded by the browser and run on-device.
Can the result be used as a final moderation decision?
Not by itself. NSFW models can produce false positives or miss content, so use this as a pre-screening signal with human review for publishing, moderation, or compliance.
Why is this model larger?
It uses a ViT classifier with about 56.8 MB of weights. First load can be slower, later loads may use browser cache, and WASM is used when WebNN is unsupported.

Related tools

  • Image to Base64

    Convert image files to Base64 strings

  • Image Upscaling and Restoration

    A free-forever, unlimited Real-ESRGAN 4x image upscaling and restoration tool with general upscaling, photo enhancement, compression repair, line-art enhancement, and pre-upscale sharpening. Images are processed locally and never uploaded.

  • AI Background Remover

    Remove portrait backgrounds locally in the browser with the open-source MODNet ONNX model, preferring WebNN NPU/GPU while never uploading images

  • AI Image Classifier

    Identify image content locally in the browser with the open-source MobileNetV2 ONNX model, preferring WebNN NPU/GPU while never uploading images

  • AI Image Tag Generator

    Generate image tags for free in the browser with the open-source MobileNetV2 ONNX model, preferring WebNN while never uploading images

  • AI Deepfake Detector

    Detect possible deepfake or AI-generated images for free in the browser with an open-source ONNX model, preferring WebNN while never uploading images

  • AI Generated Image Detector

    Detect possible AI-generated or fake images for free in the browser with an open-source ONNX model, preferring WebNN while never uploading images

  • AI Food Classifier

    Identify food photos for free in the browser with the open-source Food-101 Swin ONNX model, preferring WebNN while never uploading images

  • AI Face Emotion Detector

    Detect facial expressions for free in the browser with an open-source ViT ONNX model, preferring WebNN while never uploading images

  • Browser OCR

    A local PP-OCRv5 mobile text detection and recognition entry for screenshots, receipts, IDs, and form images

  • Document Layout and Table Detection

    A PP-DocLayout browser-side document structure detection entry for titles, paragraphs, tables, and figures

  • Monocular Depth Estimation

    A Depth Anything V2 Small ONNX/WebNN browser tool for generating relative depth maps

  • SVG Preview

    Preview and edit SVG code online

  • Image Compressor

    Compress images right in your browser — adjust quality, cap the max width, and pick the output format. Nothing is uploaded.

  • Image Format Converter

    Convert images between PNG, JPEG, and WebP directly in your browser. Free and private.

  • Image Resizer

    Resize image width and height in your browser with aspect-ratio lock, quality control, and format export.