User Generated Content can be hard to interpret automatically. Individual writing styles can differ substantially, alongside other forces of language variation. Take advantage of a robust system, ready to deal with a variety of noisy text data.
We adhere to the latest standards in bias evaluation and mitigation to protect frequently targeted social groups and foster unhindered dicussion on controversial topics. It’s about how you discuss, not what you discuss.
Text classification using Hexis API default models achieves over 80% accuracy on completely unseen data - proven on large industry datasets. Given enough training data, custom models can easily go further.
Integration is as easy as generating an access key and integrating a piece of code into your application. You find examples for various programming languages in the Documentation. The service scales automatically according your needs.
We follow a Zero Data Retention Policy. Your data never touches a server harddisk and is protected by hardened connection security. All services are hosted on dedicated infrastructure and operated from a secure datacenter in Austria.
Instant sign up. No credit card required.
Hexis API is to be used programmatically, i.e. through the use of a programming language or rather, one of its convenient standard libraries (see the Documentation for more).
The API consists of a REST endpoint secured by SSL. Authentication happens via a JWT bearer token that can be regenerated at any time in the account page. Data submitted for classification over this encrypted channel is directly passed to the classification model which returns a simple number, its score.
The data itself never gets saved or copied. Furthermore, since our services operate from within Austria, all data are protected by the GDPR and Strong Customer Protection directives of the European Union.
No problem. All detection features can be tuned in a custom model. Drop us a line.
A rigorous data curation process needs to be in place in order to train robust models, so data are routinely screened for various kinds of sample bias. In order to prevent social biases, we have developed a testing suite for probing erroneous representations of e.g. identity terms and use this information to debias our models.
We utilize Stripe, an industry-leading, PCI-compliant payment service, for our billing needs. We never see or store your confidential payment information, and all transactions are secured by SSL.
We accept all major credit cards. And soon bank transfers via SEPA.
Absolutely not, you may cancel anytime.