Ai image Detection
Our image detection feature enables AI-generated image detection by verifying image metadata and using a machine learning system trained to differentiate between human and AI-generated images.
curl --request POST \
--url https://api.gowinston.ai/v2/image-detection \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '{
"url": "<string>",
"version": "<string>"
}'{
"score": 123,
"human_probability": 123,
"ai_probability": 123,
"version": "<string>",
"mime_type": "<string>",
"c2pa": {
"active_manifest": {
"vendor": "<string>",
"claim_generator": "<string>",
"title": "<string>",
"credentials": {},
"thumbnail": {},
"ingredients": [
"<any>"
],
"ingredient_paths": [
"<any>"
],
"assertions": {},
"alg": "<string>",
"ta_url": "<string>",
"private_key": "<string>",
"sign_cert": "<string>"
},
"manifests": {
"<manifest_key>": {
"vendor": "<string>",
"claim_generator": "<string>",
"title": "<string>",
"credentials": {},
"thumbnail": {},
"ingredients": [
"<any>"
],
"ingredient_paths": [
"<any>"
],
"assertions": {},
"alg": "<string>",
"ta_url": "<string>",
"private_key": "<string>",
"sign_cert": "<string>"
}
}
},
"exif": {
"DateCreated": "<string>",
"Description": "<string>",
"Headline": "<string>",
"MaxAvailWidth": "<string>",
"MaxAvailHeight": "<string>",
"ImageRating": "<string>",
"DescriptionWriter": "<string>",
"Genre": "<string>",
"RightsUsageTerms": "<string>",
"SceneCode": "<string>",
"AdditionalModelInformation": "<string>",
"Creator": "<string>",
"CreditLine": "<string>",
"CopyrightNotice": "<string>",
"ImageSupplier": "<string>",
"Locationcreated": "<string>",
"MinorModelAgeDisclosure": "<string>",
"WebStatementofRights": "<string>",
"ModelAge": "<string>",
"ImageSupplierImageID": "<string>",
"LinkedEncodedRightsExpression": "<string>",
"DataMining": "<string>",
"Artwork": "<string>",
"ImageRegistryEntry": "<string>",
"City": "<string>",
"CodeofOrganisationFeaturedintheImage": "<string>",
"EventIdentifier": "<string>",
"EventName": "<string>",
"Contributor": "<string>",
"CopyrightOwner": "<string>",
"Country": "<string>",
"DigitalSourceType": "<string>",
"PropertyReleaseId": "<string>",
"PropertyReleaseStatus": "<string>",
"NameofOrganisationFeaturedintheImage": "<string>",
"Source": "<string>",
"DigitalImageGUID": "<string>",
"Licensor": "<string>",
"keywords": [
"<any>"
],
"ImageCreator": {},
"Instructions": "<string>",
"Title": "<string>"
},
"ai_watermark_detected": true,
"ai_watermark_issuers": {},
"credits_used": 123,
"credits_remaining": 123
}Authorizations
Bearer authentication header of the form Bearer <token>, where <token> is your auth token.
Body
Specifies the URL of the image to scan. The URL must be valid, publicly accessible, and point to an image in one of the following formats: JPG, JPEG, PNG, or WEBP. The image must have a minimum resolution of 256x256 pixels.
The model version to use. Our latest and most accurate version is "2". Using "latest" will make sure you are always using the latest version. Options: 2, 1, latest. Default: 2
Response
The human score that the image is likely to be generated by a human. A low score means our system believes that the image is generated by AI, and a higher score means that the system believes that the image is generated by a Human. The score is between 0 and 100. 0 for AI and 100 for Human.
The human probability that the image is likely to be generated by a human. The probability is between 0 and 1.
The AI probability that the image is likely to be generated by an AI. The probability is between 0 and 1.
The model version used to generate the prediction.
The mime type of the image.
The Content Provenance and Authenticity (C2PA) metadata of the image. For more information about what is C2PA, please visit https://c2pa.org/
The last manifest in the list of manifests which is the one with the set of content bindings that are able to be validated.
Typically an Internet domain name (without the TLD) for the vendor (i.e. adobe, nytimes). If provided this will be used as a prefix on generated manifest labels.
A UserAgent string that will let a user know what software/hardware/system produced this Manifest - names should not contain spaces (defaults to c2patool).
A human-readable string to be displayed as the title for this Manifest (defaults to the name of the file this manifest was embedded in).
An array of W3C verifiable credentials objects defined in the c2pa assertion specification. Section 7.
An object with an identifier field with a file path, and a format with the mime type of that file.
Ingredients that were used to modify the asset referenced by this Manifest (if any).
File paths to assets that were used to modify the asset referenced by this Manifest (if any). This may be a JSON Ingredient definition file.
Objects with label, and data - standard c2pa labels must match values as defined in the c2pa assertion specification.
Signing algorithm: one of [ ps256 | ps384 | ps512 | es256 | es384 | es512 | ed25519]. Defaults to es256.
A URL to an RFC3161 compliant Time Stamp Authority. If missing there will no secure timestamp.
File path to a private key file.
File path to signing cert file.
A collection of manifests attached to an asset. Each manifest contains information about the provenance of the asset. Creating or editing an asset using a C2PA-compliant device or tool (for example Adobe Photoshop) adds a new manifest to the manifest store. See more information at the C2PA website (https://c2pa.org/).
Typically an Internet domain name (without the TLD) for the vendor (i.e. adobe, nytimes). If provided this will be used as a prefix on generated manifest labels.
A UserAgent string that will let a user know what software/hardware/system produced this Manifest - names should not contain spaces (defaults to c2patool).
A human-readable string to be displayed as the title for this Manifest (defaults to the name of the file this manifest was embedded in).
An array of W3C verifiable credentials objects defined in the c2pa assertion specification. Section 7.
An object with an identifier field with a file path, and a format with the mime type of that file.
Ingredients that were used to modify the asset referenced by this Manifest (if any).
File paths to assets that were used to modify the asset referenced by this Manifest (if any). This may be a JSON Ingredient definition file.
Objects with label, and data - standard c2pa labels must match values as defined in the c2pa assertion specification.
Signing algorithm: one of [ ps256 | ps384 | ps512 | es256 | es384 | es512 | ed25519]. Defaults to es256.
A URL to an RFC3161 compliant Time Stamp Authority. If missing there will no secure timestamp.
File path to a private key file.
File path to signing cert file.
The exif metadata of the image. We currently retrieve the XMP, EXIF, and IPTC metadata of the image to help us determine if the image was AI generated. For more information about the IPTC metadata, please visit https://iptc.org/
Designates the date and optionally the time the content of the image was created rather than the date of the creation of the digital representation.
A textual description, including captions, of the image.
A brief synopsis of the caption. Headline is not the same as Title.
The maximum available width in pixels of the original photo from which this photo has been derived by downsizing.
The maximum available height in pixels of the original photo from which this photo has been derived by downsizing.
Rating of the image by its user or supplier
Identifier or the name of the person(s) involved in writing, editing or correcting the Description, Alt Text (Accessibility), or Extended Description (Accessibility) of the image.
Artistic, style, journalistic, product or other genre(s) of the image (expressed by a term from any Controlled Vocabulary)
The licensing parameters of the image expressed in free-text.
Describes the scene of a photo content. Specifies one ore more terms from the IPTC Scene-NewsCodes. Each Scene is represented as a string of 6 digits in an unordered list.
Information about the ethnicity and other facets of the model(s) in a model-released image.
Contains the name of the photographer, but in cases where the photographer should not be identified the name of a company or organisation may be appropriate.
The credit to person(s) and/or organisation(s) required by the supplier of the image to be used when published. This is a free-text field.
Contains any necessary copyright notice for claiming the intellectual property for this photograph and should identify the current owner of the copyright for the photograph. Other entities like the creator of the photograph may be added in the corresponding field. Notes on usage rights should be provided in Rights usage terms.
Identifies the most recent supplier of the image, who is not necessarily its owner or creator.
The location the photo was taken.
Age of the youngest model pictured in the image, at the time that the image was made.
URL referencing a web resource providing a statement of the copyright ownership and usage rights of the image.
Age of the human model(s) at the time this image was taken in a model released image.
Optional identifier assigned by the Image Supplier to the image.
A linked rights expression using any rights expression language.
Data mining prohibition or permission, optionally with constraints.
A set of metadata about artwork or an object in the image
Both a Registry Item Id and a Registry Organisation Id to record any registration of this digital image with a registry.
Name of the city of the location shown in the image. This element is at the third level of a top-down geographical hierarchy.
Code from a controlled vocabulary for identifying the organisation or company which is featured in the image.
Identifier(s) of the specific event at which the photo was taken
Names or describes the specific event at which the photo was taken.
Party or parties (person or organisation) which contributed to the image, refinement by the role attribute.
Owner or owners of the copyright in the licensed image.
Full name of the country of the location shown in the image. This element is at the top/first level of a top-down geographical hierarchy. The full name should be expressed as a verbal name and not as a code, a code should go to the element CountryCode
Name of the organisation or company which is featured in the image.
Identifier(s) of a Property Release document.
Summarises the availability and scope of property releases authorising usage of the properties appearing in the photograph.
Name of the organisation or company which is featured in the image.
The name of a person or party who has a role in the content supply chain. This could be an agency, a member of an agency, an individual or a combination. Source could be different from Creator and from the entities in the Copyright Notice.
Globally unique identifier for this digital image. It is created and applied by the creator of the digital image at the time of its creation . This value shall not be changed after that time.
A person or company that should be contacted to obtain a licence for using the item or who has licensed the item.
Keywords to express the subject of the image. Keywords may be free text and don't have to be taken from a controlled vocabulary. Codes from the controlled vocabulary IPTC Subject NewsCodes must go to the Subject Code field.
Creator or creators of the image
Any number of instructions from the provider or creator to the receiver of the image
A shorthand reference for the digital image. Title provides a short human readable name which can be a text and/or numeric reference. It is not the same as Headline.
Whether an AI watermark was detected in the image.
The list of AI watermark issuers detected in the image.
The credits_used field represents the number of credits consumed for processing your request. Each image that is processed by the API consumes 300 credits.
The credits_remaining field shows how many credits you have left in your account after your request has been processed.
curl --request POST \
--url https://api.gowinston.ai/v2/image-detection \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '{
"url": "<string>",
"version": "<string>"
}'{
"score": 123,
"human_probability": 123,
"ai_probability": 123,
"version": "<string>",
"mime_type": "<string>",
"c2pa": {
"active_manifest": {
"vendor": "<string>",
"claim_generator": "<string>",
"title": "<string>",
"credentials": {},
"thumbnail": {},
"ingredients": [
"<any>"
],
"ingredient_paths": [
"<any>"
],
"assertions": {},
"alg": "<string>",
"ta_url": "<string>",
"private_key": "<string>",
"sign_cert": "<string>"
},
"manifests": {
"<manifest_key>": {
"vendor": "<string>",
"claim_generator": "<string>",
"title": "<string>",
"credentials": {},
"thumbnail": {},
"ingredients": [
"<any>"
],
"ingredient_paths": [
"<any>"
],
"assertions": {},
"alg": "<string>",
"ta_url": "<string>",
"private_key": "<string>",
"sign_cert": "<string>"
}
}
},
"exif": {
"DateCreated": "<string>",
"Description": "<string>",
"Headline": "<string>",
"MaxAvailWidth": "<string>",
"MaxAvailHeight": "<string>",
"ImageRating": "<string>",
"DescriptionWriter": "<string>",
"Genre": "<string>",
"RightsUsageTerms": "<string>",
"SceneCode": "<string>",
"AdditionalModelInformation": "<string>",
"Creator": "<string>",
"CreditLine": "<string>",
"CopyrightNotice": "<string>",
"ImageSupplier": "<string>",
"Locationcreated": "<string>",
"MinorModelAgeDisclosure": "<string>",
"WebStatementofRights": "<string>",
"ModelAge": "<string>",
"ImageSupplierImageID": "<string>",
"LinkedEncodedRightsExpression": "<string>",
"DataMining": "<string>",
"Artwork": "<string>",
"ImageRegistryEntry": "<string>",
"City": "<string>",
"CodeofOrganisationFeaturedintheImage": "<string>",
"EventIdentifier": "<string>",
"EventName": "<string>",
"Contributor": "<string>",
"CopyrightOwner": "<string>",
"Country": "<string>",
"DigitalSourceType": "<string>",
"PropertyReleaseId": "<string>",
"PropertyReleaseStatus": "<string>",
"NameofOrganisationFeaturedintheImage": "<string>",
"Source": "<string>",
"DigitalImageGUID": "<string>",
"Licensor": "<string>",
"keywords": [
"<any>"
],
"ImageCreator": {},
"Instructions": "<string>",
"Title": "<string>"
},
"ai_watermark_detected": true,
"ai_watermark_issuers": {},
"credits_used": 123,
"credits_remaining": 123
}