The Babel Machine is a no-code artificial intelligence tool developed by poltextLAB for automated labeling and analysis of large-scale structured text corpora. The platform supports over 30 languages—including numerous Central and Eastern European languages such as Hungarian, Polish, Czech, and Slovak—and offers 14 different analytical tasks. These include sentiment analysis, emotion detection, policy topic classification (on Comparative Agendas Project codebook), named entity recognition (NER), and identification of illiberal political frames (ILLFRAMES). The system is built on XLM-RoBERTa multilingual transformer models, which enable cross-lingual transfer learning: a model trained on English texts can be applied to other languages with minimal target-language data.

The platform is used by hundreds of researchers, analysts, and journalists worldwide, and its models have been downloaded thousands of times from the HuggingFace model repository. The Babel Machine operates on GDPR-compliant European infrastructure, with all processing conducted on servers located within the European Union. The methodology has been validated through a peer-reviewed scientific publication (Sebők et al. 2024, Social Science Computer Review), ensuring the reliability and reproducibility of results. The tool is particularly valuable for researchers and professionals who wish to conduct large-scale, multilingual text analysis in areas such as political communication, media research, or policy analysis without requiring programming expertise.