Resources

Message

A Message object contains all the data related to an analysis request.

AttributeTypeRequiredDescription
textStringtrueThe text message to analyze.
⚠️  Max length is limited to 1000 characters.
contextContextfalseSome context information that will be used on the dashboard to improve filtering.
referenceStringfalseA custom unique identifier for this Message.
Auto-generated by default.
languageStringfalseThe message language (ISO 639-1).
If not set, our system will try to auto-detect the language.

Context

This resource allows you to benefit from advanced analyzer services.

AttributeTypeRequiredDescription
topLevelReferenceStringfalseAllows you to link this message to the top level message.
e.g first/original post.
parentReferenceStringfalseAllows you to link this message to a parent message.
fromAuthorfalseAllows you to add any user sender details.
toAuthorfalseAllows you to add any user receiver details.
contentContentfalseAllows you to add any content details.

Author

An Author resource contains informations regarding a user involved in the Message (sender or receiver).

AttributeTypeRequiredDescription
typeAuthorTypetrueAuthor type.
dataAuthorDetailsfalseThe author's details.
Not applicable for ANONYMOUS type.

Author details

AttributeTypeRequiredDescription
identifierStringtrueThe author's unique identifier.
profilePictureURLStringfalseThe author's profile picture URL.
usernameStringfalseThe author's username.
birthdateDatefalseThe author's birthdate (ISO 8601 format).
genderGenderfalseThe author's gender.

Content

A Content resource contains information regarding the content involved in the Message.

AttributeTypeRequiredDescription
typeContentTypetrueContent type.
dataContentDetailsfalseThe content's details.
Not applicable for UNKNOWN type.

Content details

AttributeTypeRequiredDescription
identifierStringtrueThe content's unique identifier.
titleStringfalseThe content's title.

Analysis

An Analysis object contains all the data related to an analysis response.

AttributeTypeDescription
textStringThe analyzed text.
referenceStringYour custom reference.
typeTypeDecides the nature of the content.
severitySeverityDetects the severity of the content.
recommendedActionRecommendedActionThe action recommended by our system for the analyzed text.
metaMetaMeta information.
languageStringString formatted language (ISO639-1)
analyzedAtStringString formatted date (ISO8601).

Meta

AttributeTypeDescription
directedAtDirectedAtDetermines who the content is directed at.
classifications[Classification]Identifies the category of toxic content.

Enum types

Author types

ValueDescription
AUTHORThe user is identifiable.
See AuthorDetails
ANONYMOUSThe user is anonymous.

Content types

ValueDescription
VIDEOVideo content.
PHOTOPhoto content.
AUDIOAudio content.
LINKLink content.
TEXTText content.
OTHEROther content.

Gender

ValueDescription
MALEMale gender
FEMALEFemale gender
OTHEROther gender
NAUndefined gender

Type

ValueDescription
HATEFULContent analyzed is hateful.
e.g: You are ugly.
NEUTRALContent analyzed is not hateful.
e.g: I feel good.
SUPPORTIVEContent analyzed is supportive.
e.g: You're pretty.
NOISEContent analyzed is "unnecessary noise" (spam, ads, scam...).
e.g: Please subscribe to my channel.

Severity

ValueDescription
NONEContent analyzed is neutral.
e.g: Please subscribe to my channel.
LOWContent analyzed has low severity.
e.g: Einstein is an idiot.
MEDIUMContent analyzed has medium severity.
e.g: You are useless.
HIGHContent analyzed has high severity.
e.g: You're a dumb bitch.
VERY_HIGHContent analyzed has very high severity.
e.g: You're a big stupid assholes.

Recommended Action

ValueDescription
REMOVEContent analyzed should be removed.
KEEPContent analyzed should be kept.
WATCHContent analyzed has been detected as a very low severity and should be double-checked.

Directed At

ValuesDescription
USERContent analyzed is directed at a user (“you”).
e.g: You are an idiot.
USER_FAMILYContent analyzed is directed at the user’s family (mom, dad, brother etc.).
e.g: Your brother is an idiot.
SINGLE_PERSONContent analyzed is directed at a third party (he, she, name).
e.g: Einstein is an idiot.
GROUPContent analyzed is directed at a group of people (like a minority).
e.g: Einstein and scientists are idiots.
EVERYONEContent analyzed is directed at everyone.
e.g: we need to slaughter yall.
AUTHOR_OF_COMMENTContent analyzed is directed at the author of the comment themselves.
e.g: I'm so stupid.
NO_ONEContent analyzed is not directed at anyone.
e.g: It’s not very nice when you call someone a whore...
HATERSContent analyzed is directed at the "haters".
e.g: Fuck the haters, you are beautiful.

Classifications

ValuesDescription
INSULTText that offends, that hurts someone’s dignity.
e.g: You are an idiot.
THREATText indicating an intention to harm someone, to make them do something against their will.
e.g: I'm going to kill you.
TROLLINGText that aims to create controversy no matter the topic of the conversation; that looks to pollute in particular through disrupting the discussion space.
e.g: Your title is a clickbait, i’m unsubscribing...
BODY_SHAMINGText that subjects someone to criticism, ridicule or mockery for supposed bodily faults or imperfections.
e.g: You are ugly.
RACISMText depicting a belief that race, skin color, language, nationality, national or ethnical origins justifies contempt for a person or group of people.
e.g: Go back to your country.
HATREDText demonstrating an extreme intolerance towards a person or group of people.
e.g: I hate you.
HOMOPHOBIAText describing a prejudice-based intolerance towards homosexuality, gays, lesbians, bisexuals.
e.g: Y'all are big faggots.
SEXUAL_HARASSMENTText depicting harassment of sexual nature.
e.g: You're a hottie.
MORAL_HARASSMENTText meant to annoy someone, to intimidate or humiliate them, with the purpose of degrading someone’s life through targeting their physical or mental health.
e.g: U should kill urself
MISOGYNYText expressing hatred of, or prejudice against women or girls. It relates to sexism, promoting an inferior status for women and rewarding those who accept it.
e.g: Go back to the kitchen, women.
SUPPORTIVEText identified as supportive.
e.g: You are so nice.
NEUTRALText identified is not hateful.
e.g: Hey, how are you ?
ADSText identified as ads, promotion.
e.g: Please subscribe to my channel.
USELESSText is useless, adds no value to the conversation.
e.g: FIRST !!!!
LINKText containing links.
e.g: Hey, do you want to win an iPhone ? Go to free-iphone.tk
SCAMText indicating a scam.
e.g: Do you want to win a lot of money easily, let’s DM
SPAMText that is spam.

All fields respect the General Data Protection Regulation (GDPR).