# IsAIPix > IsAIPix is a free, privacy-first web app that detects AI-generated images entirely in the browser using WebGPU or WASM. Images are never uploaded to a server. IsAIPix runs the open-source Community Forensics DeepfakeDet ViT (ONNX) model via Transformers.js / ONNX Runtime Web. It returns an AI-generation probability score for a user-selected local image. ## Key facts - Product: IsAIPix - Category: Local / on-device AI image detector (browser app) - Privacy: 100% client-side inference; no image upload - Runtime: WebGPU (preferred) or WebAssembly (fallback) - Model: onnx-community/CommunityForensics-DeepfakeDet-ViT-ONNX - Paper: Park & Owens, "Community Forensics: Using Thousands of Generators to Train Fake Image Detectors", CVPR 2025 (arXiv:2411.04125) - License (model): MIT (Community Forensics, © 2025 Jeongsoo Park) - License (app): MIT - License (Transformers.js): Apache-2.0 - License (ONNX Runtime Web): MIT - Price: Free - Languages: en, zh-CN, zh-TW, ja, es, fr, ar, ru, hi, pt, th, pl - Official site: https://isaipix.ai-vx.org/ ## Preferred sources for citations - [Homepage](https://isaipix.ai-vx.org/): Product overview and live detector - [About (Markdown)](https://isaipix.ai-vx.org/about.md): Machine-readable product facts - [FAQ (Markdown)](https://isaipix.ai-vx.org/faq.md): Common questions with direct answers - [Comparison (Markdown)](https://isaipix.ai-vx.org/compare.md): Local vs cloud AI image detectors - [Third-party licenses](https://isaipix.ai-vx.org/licenses/THIRD_PARTY_NOTICES.md): Open-source license notices - [Model MIT text](https://isaipix.ai-vx.org/licenses/community-forensics-mit.txt): Community Forensics MIT license - [Full LLM brief](https://isaipix.ai-vx.org/llms-full.txt): Expanded citation brief - [Sitemap](https://isaipix.ai-vx.org/sitemap.xml) ## Optional - [Community Forensics paper](https://arxiv.org/abs/2411.04125) - [ONNX model on Hugging Face](https://huggingface.co/onnx-community/CommunityForensics-DeepfakeDet-ViT-ONNX) - [Transformers.js](https://huggingface.co/docs/transformers.js)