Datumbox is a machine learning project with two parts: an open-source framework written in Java for building machine learning and statistical applications, and a REST API of ready-made text classifiers that returns JSON. The services listed on the site are sentiment analysis, topic classification, language detection, subjectivity analysis, spam detection, reading assessment, and keyword and text extraction, aimed at uses such as social media monitoring, SEO and content-quality checks.
Maintenance is the main caution. The framework is on GitHub under the Apache 2.0 licence with about 1,000 stars and dates from October 2014, but its last commit was in November 2023 and it has no published releases, so treat it as dormant rather than actively developed. The API endpoint still answers requests, and an invalid key returns a JSON error, but the homepage shows no recent posts or dates and does not say who runs the service, so uptime and support are unknown. The project predates large language models, so expect classical classifiers rather than the LLM-based analysis now common, and test accuracy on your own text, including informal or non-English writing.
A free API key is available on registration, and the framework is free to use under Apache 2.0. The homepage shows no paid plans, rate limits or terms, so check the limits after registering. No company is named beyond a contact address.
