Frequently asked questions

PixGuard runs each image through a multi-stage pipeline. First, AI models check for AI generated content (Midjourney, DALL-E, Stable Diffusion). Then computer vision detectors scan for visible watermarks (Getty, Shutterstock, Adobe Stock). Embedding models compare against a database of 600,000 known stock images. OCR extracts hidden text. Metadata analysis reads EXIF, IPTC, XMP, and C2PA Content Credentials. Each signal contributes to a final risk score from 0 to 100.

On our internal test set, PixGuard catches 94 percent of true copyright infringement cases with a 3 percent false positive rate. Accuracy varies by image type. Visible watermarks are detected with 98 percent accuracy. Invisible watermarks vary by encoding strength. AI generated detection is 91 percent accurate on photorealistic content. We continuously improve detection by collecting feedback on flagged images.

Yes. Enter any website URL and PixGuard crawls it, extracts all images, and scans each one. Free tier scans up to 10 images per site, Starter scans 50, Pro scans 500, and Business scans up to 10,000 images. Scans happen in the background and you get an email when complete.

PixGuard pricing is measured in image scans. Free users get ~30 image scans on signup with no card required. Paid plans start at $9.99 per month for ~50 image scans. Business is $79.99 per month for ~650 image scans. Source attribution is included on every paid scan. Previously scanned images are always free to rescan.

Google reverse image search finds copies of an image online. PixGuard tells you if an image is at risk of copyright infringement. We detect watermarks, metadata flags, AI generation, and stock database matches. Google does not score risk or identify the rights holder. Use Google to find where an image appears, use PixGuard to know if you can safely use it.

TinEye is excellent at finding exact copies of an image online. PixGuard is a full copyright detection pipeline that catches risks TinEye cannot see. Watermark text via OCR. AI generated images via ViT classifiers. Removed watermarks via Error Level Analysis. Stock database fingerprints across 600,000 images. PixGuard also has a free tier with ~30 image scans, while TinEye charges $200 per month.

Yes. We use a Vision Transformer model fine tuned to detect AI generated content from Midjourney, DALL-E, Stable Diffusion, Adobe Firefly, and Flux. We also analyze pixel level noise patterns since real cameras produce specific sensor noise that AI tools cannot replicate. Plus we read C2PA Content Credentials when present. Detection happens in roughly 60 milliseconds per image.

Yes. Images you upload are processed in memory and not stored on our servers. We cache only the perceptual hash and analysis result, not the original image. Cached results expire after 30 days. We never share your scan history with third parties. Source attribution lookups go through Google Vision and SerpAPI. Read our Privacy Policy for details.

Yes. We offer a Chrome extension that scans any image you right click. A WordPress plugin that audits your media library and runs weekly scans. A REST API for Pro and Business plans. SSE streaming for real time results. And a public report URL you can share with team members or legal counsel.

You get a risk score from 0 to 100 plus detailed signals showing exactly what was detected. For example, "OCR detected getty images watermark" or "Visual match to known Shutterstock asset (87 percent confidence)". You also get an action recommendation. Replace the image, license it properly, or proceed if confidence is low. Pro plans include a generated DMCA template if you need to challenge a takedown.

Cache hits return in under 15 milliseconds. Fresh scans typically complete in 300 to 500 milliseconds for the analysis. Source attribution adds 15 to 20 seconds when enabled. Website scans depend on size, typically 30 seconds to 2 minutes for a small site. Results stream progressively so you see partial results immediately.

Yes, on Pro and Business plans. Upload up to 100 images at once via the dashboard or use our API to submit batches programmatically. WordPress plugin runs automatic batch scans on your media library. Free and Starter tiers can scan one image at a time.

C2PA stands for Coalition for Content Provenance and Authenticity. It is a cryptographic signature embedded in images by Adobe, Microsoft, OpenAI, and major camera manufacturers. C2PA proves who created an image, when, and what tools were used. PixGuard reads C2PA chains to verify image authenticity instantly without running our full pipeline.

Yes. Cancel from your settings page anytime. Your subscription continues until the end of your billing period, then drops to free tier. You keep any unused image packs permanently. Subscription allowances expire at the end of each billing period.

You get ~30 free image scans when you sign up (full detection on your first 3 scans, then lighter detection afterward). No credit card required. Use it to scan your website, test on sample images, or evaluate accuracy before upgrading.

Each result has a thumbs up and thumbs down feedback button. Your feedback trains our detection models to be more accurate. False positives on common safe images get fixed quickly. You can also contact [email protected] with the image URL and we will investigate within 24 hours.