Transcription Portal Presented at CLARIN Café: Speech-to-Text Made Accessible for Researchers
Online session on 29 October drew 35 participants and showcased improvements shaped by the 2nd ATRIUM Research Forum: Transcription Portal.
The CLARIN Café hosted an online session dedicated to the TranscriptionPortal , a web service that turns spoken language into text. The event was marked by lively discussion and active exchange throughout.
Making spoken data usable
Spoken language is a key source of information in many research areas, in fieldwork and in everyday life. The TranscriptionPortal makes this information accessible through a simple three-step process: users upload their audio, select the input and output languages, and start processing. The platform currently supports five languages: English, Dutch, Italian, German and Greek.
The typical workflow combines automatic speech recognition with optional manual validation, translation and summarization. The services are preconfigured, so that only a small number of options need to be set, which makes the tool well suited to non-technical users. Results are provided as downloadable text files containing the transcribed and translated text, either in full length or as a summary.
The TranscriptionPortal is free for academic institutions. Privacy is a core design principle: data is stored only during processing, all communication is encrypted, and only trusted academic service providers are accessed.
Improvements driven by community feedback
A highlight of the session was the presentation of recent modifications and enhancements to the portal, which build directly on input gathered at the ATRIUM Research Forum on 25 November 2025. The Café offered participants the opportunity to see how this feedback has been translated into concrete improvements, and to contribute their own ideas for future development.
The session was led by Christoph Draxler, head of the Bavarian Archive for Speech Signals at LMU Munich, and Henk van den Heuvel, senior researcher and Head of the Humanities Lab at Radboud University and member of CLARIN’s Board of Directors. The programme included an introduction, a live demonstration, an overview of new features, a discussion of use cases, a collection of ideas for new features and improvements, and a closing comparison with other available solutions.
About the TranscriptionPortal
The TranscriptionPortal is operated by the Bavarian Archive for Speech Signals (BAS) in close collaboration with the Centre for Language and Speech Technology (CLST) , Radboud University and is supported by CLARIN . It was successfully evaluated in a usability test in November 2025.
About the CLARIN Café
The CLARIN Café is a series of online sessions on tools, resources and topics relevant to researchers in linguistics and language sciences, digital humanities, language technology and NLP, social sciences, speech technology, and language teaching. A full overview of planned and previous sessions is available on the CLARIN Café page .