One result = one image or one page of a PDF. A 12-page PDF uses 12 results - there's no per-document cap.
Nothing breaks. Paid plans bill overage at your plan's flat per-result rate, with a running total visible in your dashboard from 80% of quota onward.
No. Free includes 25 results/month with no credit card and no time limit.
Around 100,000 results/month (roughly $1,000 on Business) our team reaches out - or contact sales directly for volume pricing, SLAs, and data residency.
No. The plan price is a monthly minimum that includes your results at that plan's rate. 40,000 results on Business is $399 total, not $399 plus per-result.
Per organization - there are no seat fees. Invite unlimited team members under Settings → Teams & Roles and everyone draws from the same monthly result pool.
No. Results reset at the start of each billing cycle and don't carry forward. If you go over, overage bills automatically at your plan's per-result rate - moving up a tier lowers that rate.
No. Paid plans are month-to-month and you can move between tiers at any time. Annual prepay saves 20%.
Yes — Free and paid plans include API access. Generate your key in the dashboard and start detecting.
No. Everything runs in the browser - single uploads, bulk ZIP uploads, results, heatmaps and reports. The API is optional.
Upload a ZIP rather than a folder, in batches of 500 images or fewer for the most reliable results. Budget roughly 1-2 seconds per image.
JPG, PNG, JPEG, TIF and WEBP, up to 10 MB per file. Native resolution gives the most accurate result.
We cover the full range of manipulation: fully AI-generated images, AI-assisted edits, traditional digital retouching, and face swaps. Localized edits are where we stand apart. Most detectors are built around whole-image scanning and lose the signal when only a region has been altered - a swapped face, a changed digit, a removed object. We analyze at the region level, so a partial edit to an otherwise authentic photo gets surfaced rather than averaged away against the untouched pixels. For very small edits, cropping to the region in question and rescanning gives the cleanest read.
The pipeline is built to survive the handling images actually go through in the wild - platform re-encoding, JPEG compression, resizing, and the generational loss of a file that's been forwarded or re-exported. We also test against deliberate laundering: noise injection, filtering, and the other techniques used to push a manipulated image past detection. Heavy degradation still costs you. A screenshot of a screenshot, or a thumbnail recompressed to a fraction of its original size, discards pixel-level signal that no detector can recover.
No. Every result includes the probability score, classification, confidence score, detailed reasoning and heatmap at no additional cost.
Images are retained by default. Zero data retention (ZDR) is available on Business and above. With ZDR, your submissions and customers' data are immediately discarded once detection is complete and never used for training.
Yes. We'll sign a DPA on Business plans and above - contact sales to start one.
Yes. Regional processing is available on Enterprise and above. It runs on a dedicated endpoint and takes about a week to set up, with no impact on latency or throughput.