Trustlevel API call

API Endpoint package

Text Analytics

API endpoint

https://app.trustservista.com/api/rest/v2/trustlevel

Supported Languages

English (native).

Russian, Arabic, Spanish, French, Chinese (with automated translation).

Description

Given an article URL or raw text, the endpoint will return computed Content Quality Report (TrustLevel) for the content. The TrustLevel is an aggregated score of all the trustworthiness metrics computed for the content.

The TrustLevel API call will also return the following trustworthiness metrics:

• Author: named author, pseudo-author, no author
• Verified source: if it is in the TrustServista database or not
• Context Setting: the weight of factual information (people, organizations, places, time references) discovered in the article
• Polarity: very negative, negative, neutral, positive, very positive
• Chance of the article being clickbait
• The article’s TrustLevel

This endpoint is used in the TrustServista Web Dashboard when performing a URL verification and displaying the Content Quality Report card:

Read more on how TrustServista can automatically detect clickbait content.

Sample cURL call

curl -X POST \
    -H "X-TRUS-API-Key: your_api_key" \
    -H "Content-Type: application/json" \
    -H "Accept: application/json" \
    -H "Cache-Control: no-cache" \
    -d '{"contentUri": "http://www.foxnews.com/opinion/2017/03/01/president-trumps-best-speech.html" }' \
    "https://trust.servista.eu/api/rest/v2/trustlevel"

Sample JSON response

{
"trustLevel": 58,
"trustLevelType":"global",
	"trustLevelComponent": {
		"trustLevelComponents": [{
				"type": "author",
				"score": 1.0
			}, {
				"type": "entity",
				"score": 0.588
			}, {
				"type": "sentiment",
				"score": 1.0
			}
		]
	},
	"trustLevelMessage": "We found this article's author. \n This article has decent (58.80%) context setting (who/when/where). \n This article has an overall neutral sentiment (objective).  \n Patient Zero's TrustLevel is: 50% Trusted \n This article's TrustLevel is calculated using the metrics above and also takes in account Patient Zero's TrustLevel."
}

If the article was not preprocessed by TrustServista the “trustLevelType” is set to “local” and the article TrustScore is computed using just the metadata that can be extracted out of the article alone