How does pricing for text moderation work?
Pricing is done on a per-operation basis. Each request you make to our API will count as 1 or multiple operations depending on the content submitted.
For Rule-based pattern-matching algorithms, we will count 1 operation up to 1000 characters and then an extra operation for each extra 1000 characters. So if you do a request with a text that is 1800 characters long, that will count as 2 operations.
For Username moderation, each request counts as 1 operation only.
For Text Classification models, the same counting applies to each model requested. So if you analyze a text that is 1800 characters long with the general and self-harm models, that will count as 4 operations. When Rule-based algorithms and Text Classification models are combined in a single request, only the Text Classification models are counted.
Each time you perform a request to the API, the response field contains an operations key with details on the number of operations counted for this request.
Other frequent questions
- How does text moderation work?
- How do text classification models work?
- How do rule-based pattern-matching algorithms work?
- How does Sightengine's text moderation differ from keyword filtering? How do you prevent users from circumventing word filters?
- Is it possible to customize the Text Moderation API to my needs?
- What languages does text moderation work with?
- Can you automatically detect the language of the text I want to moderate?
- What is the maximum text length?