​
Moderate Text

To moderate text, you can use the text.moderate method. This method takes in a text string and returns a Promise that resolves to a TextModerationResponse object.

index.ts
const contentMod = new ContentMod({
  publicKey: "<YOUR_PUBLIC_KEY>",
  secretKey: "<YOUR_SECRET_KEY>",
});

// Simple text moderation
const response = await contentMod.text.moderate("Some bad text");

// Text moderation with additional metadata to be saved with the moderation request
const response2 = await contentMod.text.moderate("Some bad text", {
  meta: {
    userId: "1234",
  },
});

// Optionally you can pass in an webhook callback url if you want to be notified when the moderation request is completed
const response3 = await contentMod.text.moderate("Some bad text", {
  callbackUrl: "https://example.com/webhook",
});

​
Webhook Callback

If you pass in a webhook callback url, the response will only be an object with the id property.

​
Properties

​
text
string
required

Text to moderate

Example: Some bad text

​
options
object

An object of options for the moderation request

​
meta
object

Any additional metadata you want to include with the moderation request that will be saved.

Example:{"userId": "1234"}

​
callbackUrl
string

A webhook callback url that will be called when the moderation request is completed.

Example: https://example.com/webhook

​
defer
boolean

Whether to defer the moderation request. If this value is true the request will be queued and you will receive the response later in a webhook that you have set up. Or you can look it up later using the id returned.

Example: true

​
Response

The TextModerationResponse object contains the following properties:

​
id
string
required

The ID of the moderation request.

Example: 27fbdc0b-b295-46ce-93f5-81b2fc08a381

​
isSafe
boolean
required

Whether the text is considered safe.

Example: true

​
confidence
number
required

Confidence of the safety of the text from 1-100.

Example: 90

​
sentiment
string
required

Sentiment of the text, either negative, neutral or positive.

available sentiment values are:

  • negative
  • neutral (default)
  • positive

Example: positive

​
sentimentScore
number
required

Sentiment score from 1-100 (negative, neutral, positive).

Example: 90

​
riskScores
object
required

Overall risk score from 1-100.

Example: 90

​
topics
array
required

General topics the text depicts, lowercase and no punctuation.

Example: ["politics", "sports"]

​
nsfwCategories
array
required

An array of the categories of the text. The available categories are:

  • adult_content
  • suggestive_imagery
  • strong_language
  • violence_gore
  • horror_disturbing
  • alcohol
  • tobacco
  • substance_use
  • gambling
  • dating_relationship
  • medical_procedures”
  • crude_humor
  • political_content

Example: [{"category": "sexual", "severity": 90}]

​
summary
object
required

Summary of the text.

Example: {"profanity": true, "totalFlags": 1, "contentRating": "G", "language": "en"}

​
suggestedActions
object
required

Suggested actions for the text.

​
reject
boolean
required

Whether to reject the text.

Example: true

​
review
boolean
required

Whether to flag the text.

Example: true

Example: {"reject": true, "review": true}

​
content
string
required

The original text.

Example: This is the original text

​
filteredContent
string
required

The text with profanity replaced by *.

Example: This is a filtered text

​
request
object
required

The request object.

Example: {"requestId": "27fbdc0b-b295-46ce-93f5-81b2fc08a381", "timestamp": "2023-04-05T12:00:00.000Z"}

​
meta
object
required

The metadata of the request object that the user provided.

Example: {"userId": "12345"}

​
hash
string
required

The SHA-256 hash of the text. Used for content comparison purposes. Example: 93a08bd10cd367bf83f12ba6590fcaef2f8deec8a47ce4ce88a15c7dda325ffa

InstallationGet Text Moderation