Shunya Labs Releases Open Standard for Stuttering Annotation

Shunya Labs has launched Stuttering Annotation Markup Language (SAML), which it claims is the world’s first open-source standard for annotating stuttering in speech datasets, designed to improve interoperability across speech research and clinical workflows.

The Gurugram-based company said the framework was developed to address longstanding fragmentation in stuttering research, where different laboratories and institutions rely on separate annotation systems, making it difficult to compare datasets or reproduce findings.

SAML is an XML-based specification that extends W3C SSML standards and includes schema validation to ensure interoperability across tools. The framework supports annotation of dysfluencies such as repetitions, prolongations, and blocks, alongside a severity scale ranging from 0 to 8 based on duration. It also supports multi-speaker annotations for conversational and dialogue-based research.

According to the company, more than 80 million people globally are affected by stuttering, while inconsistent annotation standards have limited large-scale collaboration and reproducibility in speech research.

“SAML was created to provide a shared framework that researchers, clinicians, and developers can adopt across studies and tools,” said Founder and Chief Scientist Sourav Bandyopadhyay. “Standardisation is critical if speech research is expected to become interoperable, reproducible, and collaborative at scale.”

To encourage adoption, the company has made the specification, training resources, schema documentation, validation suite, and conversion tools freely available to the research community.