FAQ - Frequently Asked Questions

There are many possible reasons language detection can be important for your product. Grouping content by language, routing support request, knowing if and when to translate content are just a few examples. As soon as you deal with user generated content and an international audience, you will reach significantly more people by allowing them to write in their native language. Take a cooking platform as an example: grouping recipes as well as ingredients by their language allows you to only show relevant data to your users.

Detecting if a text contains profanity can be useful in many circumstances. Whenever you process text written by users which will be displayed to other users, checking for profanity can help moderate or prevent exposure. In cases where your audience might include children, profanity detection is a good first line of defense against malicious actors. Keep in mind however that profanity or bullying can take creative forms and even Linguins detection cannot guarantee 100% accuracy.

We currently only support English for detecting profanity. You can help us by telling us which other languages you would be interesed in.

There are many libraries out there providing some level of language detection. They vary a lot in the number of languages covered, performance, and most importantly accuracy. Almost all of them use a single trained ML-model which requires a certain amount of text to perform well. Maintaining a solution that combines dictionaries, n-grams (both word- and character-based), as well as trained ML-models - as Linguin does - is not trivial. The benefit: even single words and very short texts are detectable with high accuracy.

173. Here is the list.

Linguin returns 2-character ISO-639-1 language codes when available and 3-character ISO-639-2 codes otherwise. Take a look at the full language list.

Language detection can be ambiguous at times. The English word "kind" for example also means "child" in German. Based on how common a term and its language are Linguin assigns a confidence (a value between 0.00 and 1.00). It is possible for confidence scores to sum up to a number above or below 1.00.

Yes. You can detect profanity and the languages of multiple texts with one API call respectively. Keep in mind that Linguin counts the number of detections, not the number of API calls for calculating what to charge you. See the API documentation for how to use the bulk language and bulk profanity endpoints.

Linguin is forever free for up to 100 detections per day. All you need is to sign up for your free api key. For more detections you need to upgrade your account by providing a credit card. Even then you only pay when you actually use the service. 100 detections per day are still free of charge. As soon as you exceed that limit you will be charged at the end of your billing period based on the maximum detections per day in the prior month. The formula is simple: $2.50 for the first 1,000 detections + $0.15 per 1,000 detections after that.

Yes. After upgrading your account you can choose to change your daily limit of detections any time. You can even remove the limit altogether, although this is strongly discouraged.

Probably. Nobody is perfect, so we ask you to visit and fill out our form to submit mistakes. We will review every submission and try to teach Linguin to do better.