Calame: An Open Source Transcription Software

Thomas Soulas, Yves Ferstler, Valentyna Tsilinchuk, Yassine Chahdi, Catherine Lavoie, Gaëlle Laperrière, Marie-Jean Meurs
Proceedings of the The 39th Canadian Conference on Artificial Intelligence, PMLR 318:990-996, 2026.

Abstract

While research on automatic speech processing is very active, its outcomes remain mainly inaccessible to people without programming skills or expertise. Moreover, studies focus mostly on high-resource languages and conventional setups, preventing a wider adoption and social impact of these technologies. Automatic speech processing systems can be needed in a variety of use cases, such as automatic transcription of meetings, interviews, or even conferences. They can also be useful for subtitling and dictation, or to interact with voice assistants. Non-experts may rely on commercial solutions, but these typically lack modularity, o!er only partial functionalities, increase exposure to cyber threats, and impose significant financial barriers for potential users. As automatic transcription techniques improve, it becomes crucial to make these tools accessible to both the research community and the general public. To make language technology more inclusive, we released Calame, a free, open-source, and accessible software for automatic multilingual speech processing, available for both local and remote use. Its current language coverage includes English and French, with Quebec French and other low-resource languages being gradually incorporated with state-of-the-art fine-tuned models.

Cite this Paper

BibTeX
@InProceedings{pmlr-v318-soulas26a, title = {Calame: An Open Source Transcription Software}, author = {Soulas, Thomas and Ferstler, Yves and Tsilinchuk, Valentyna and Chahdi, Yassine and Lavoie, Catherine and Laperri\`{e}re, Ga{\"e}lle and Meurs, Marie-Jean}, booktitle = {Proceedings of the The 39th Canadian Conference on Artificial Intelligence}, pages = {990--996}, year = {2026}, editor = {Bouzar-Benlabiod, Lydia and Leung, Carson}, volume = {318}, series = {Proceedings of Machine Learning Research}, month = {25--29 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v318/main/assets/soulas26a/soulas26a.pdf}, url = {https://proceedings.mlr.press/v318/soulas26a.html}, abstract = {While research on automatic speech processing is very active, its outcomes remain mainly inaccessible to people without programming skills or expertise. Moreover, studies focus mostly on high-resource languages and conventional setups, preventing a wider adoption and social impact of these technologies. Automatic speech processing systems can be needed in a variety of use cases, such as automatic transcription of meetings, interviews, or even conferences. They can also be useful for subtitling and dictation, or to interact with voice assistants. Non-experts may rely on commercial solutions, but these typically lack modularity, o!er only partial functionalities, increase exposure to cyber threats, and impose significant financial barriers for potential users. As automatic transcription techniques improve, it becomes crucial to make these tools accessible to both the research community and the general public. To make language technology more inclusive, we released Calame, a free, open-source, and accessible software for automatic multilingual speech processing, available for both local and remote use. Its current language coverage includes English and French, with Quebec French and other low-resource languages being gradually incorporated with state-of-the-art fine-tuned models.} }
Endnote
%0 Conference Paper %T Calame: An Open Source Transcription Software %A Thomas Soulas %A Yves Ferstler %A Valentyna Tsilinchuk %A Yassine Chahdi %A Catherine Lavoie %A Gaëlle Laperrière %A Marie-Jean Meurs %B Proceedings of the The 39th Canadian Conference on Artificial Intelligence %C Proceedings of Machine Learning Research %D 2026 %E Lydia Bouzar-Benlabiod %E Carson Leung %F pmlr-v318-soulas26a %I PMLR %P 990--996 %U https://proceedings.mlr.press/v318/soulas26a.html %V 318 %X While research on automatic speech processing is very active, its outcomes remain mainly inaccessible to people without programming skills or expertise. Moreover, studies focus mostly on high-resource languages and conventional setups, preventing a wider adoption and social impact of these technologies. Automatic speech processing systems can be needed in a variety of use cases, such as automatic transcription of meetings, interviews, or even conferences. They can also be useful for subtitling and dictation, or to interact with voice assistants. Non-experts may rely on commercial solutions, but these typically lack modularity, o!er only partial functionalities, increase exposure to cyber threats, and impose significant financial barriers for potential users. As automatic transcription techniques improve, it becomes crucial to make these tools accessible to both the research community and the general public. To make language technology more inclusive, we released Calame, a free, open-source, and accessible software for automatic multilingual speech processing, available for both local and remote use. Its current language coverage includes English and French, with Quebec French and other low-resource languages being gradually incorporated with state-of-the-art fine-tuned models.
APA
Soulas, T., Ferstler, Y., Tsilinchuk, V., Chahdi, Y., Lavoie, C., Laperrière, G. & Meurs, M.. (2026). Calame: An Open Source Transcription Software. Proceedings of the The 39th Canadian Conference on Artificial Intelligence, in Proceedings of Machine Learning Research 318:990-996 Available from https://proceedings.mlr.press/v318/soulas26a.html.

Related Material