CAT-Talk

USING AI TO DRIVE INNOVATION AND AUTOMATION

CAT-Talk is a secure web-based platform that streamlines speech-to-text transcription and speaker-labeling (diarization), addressing a growing need for accurate, efficient, and secure documentation in healthcare and related fields. This self-service tool allows users to quickly generate, edit, and analyze transcripts, significantly reducing the burden of manual documentation. LLM-powered tools are integrated into the system allowing users to summarize transcripts or create their own custom jobs tailored to meet each project’s unique needs. Hosted on UK-owned, NIST-compliant infrastructure, CAT-Talk offers robust data security as a core advantage. Built to boost productivity, it empowers users to focus on what matters most.

Citation

A paper detailing the development and usage of this tool can be found here:

Toward Automated Clinical Transcriptions 

Please be sure to appropriately cite usage of CAT-Talk in your research.

Key Features of CAT-Talk

  • Fast, accurate transcriptions with error flags for rapid human verification.
  • Integration with a variety of LLM tools for summarization, theme extraction, and highlighting key takeaways.
  • Hosted on secure, UK-owned NIST-compliant infrastructure.
  • Customizable tools that can be tailored for specific project needs.


Learn more about CAT-Talk in this for summarization, theme extraction, and highlighting key takeaways.

tutorial video that can be tailored for specific project needs.

Vaiden Logan presented his work Toward Automated Clinical Transcriptions, at the March 2025 AMIA Informatics Summit in Pittsburgh, PA. 

Projects Supported by CAT-Talk

SpeakEZ The University of Kentucky Louie B. Nunn Center for Oral History has a backlog […]