What is Checkobot?
Checkobot checks images for signs of AI generation, AI editing and face swaps. It is aimed at people who need to assess a suspicious picture, as well as trust and safety teams that want image checks in a moderation or onboarding workflow. Results give an AI or No AI detected verdict, identify the detector that triggered it and explain the finding in plain language. A No AI detected result is not a guarantee that an image is authentic.
Checkobot Features:
Image Checks: Upload or link an image for analysis; accepted formats include JPEG, PNG, WebP, AVIF, GIF, BMP and TIFF, up to 20 MB.
Whole Image Analysis: The whole-image detector runs on every picture and looks for patterns associated with AI image generators and editors.
Face Forensics: When an image contains a clear face, a dedicated model checks for face swaps, AI face edits and synthetic faces.
Verdict Details: The result identifies the detector that fired and gives a one-line explanation, with signed Content Credentials that say AI also noted.
Shareable Results: Checkobot creates a result card on the user's device and offers a result link, with the option to include a resized image.
REST API: API results include the verdict, detector, both detector scores and face box; Growth and higher plans support batches of up to eight images per call.
Checkobot Benefits:
More Useful Evidence: Seeing what triggered a result gives reviewers context to consider alongside the image's origin, rather than leaving them with an unexplained score.
Fewer False Assumptions: The cautious verdict language helps teams avoid treating an undetected image as proven genuine or using a result to make claims about a person.
Privacy Control: Free checks are processed in memory and discarded, so people can get a second opinion without adding the image itself to stored history.
Traceable Reviews: Saved result links and recorded model versions help teams refer back to what was checked and which detection rules applied at the time.
Flexible Volume: Individuals can start with daily checks, while monthly plans offer larger allowances for regular or programmatic review.
Use Cases:
News Verification: Journalists and fact-checkers can review a viral image before publishing a story or documenting a debunk.
Account Onboarding: Trust teams can screen a signup selfie for a swapped or synthetic face before an account is opened.
Marketplace Moderation: Platforms can route user-uploaded AI images to a label, review step or block, with the triggering detector attached.
Listing Review: Property and product marketplaces can check listing photos for generated imagery or edits that may hide defects.
Claim Assessment: Insurance teams can flag claim photos that may use generated or edited images to add or enlarge damage.