AI Speech to Text Module: Real-Time Transcription for Italian Senate Speeches

I developed an advanced speech to text module powered by artificial intelligence, specifically designed for the automatic transcription of Italian Senate speeches. This module was created for the GIANO platform, a digital environment used by Nomos to manage and analyze parliamentary activities and data relevant to their interests. The solution I built can process both live streams—offering reliable real-time transcription—and pre-recorded video content already hosted on platforms such as YouTube or similar services. The ability to transcribe live video in real time is a key feature of this project, enabling immediate analysis and indexing of what is said during parliamentary sessions.

The speech to text service delivers high accuracy, even when dealing with complex or technical parliamentary language. It supports seamless integration within the GIANO ecosystem, allowing operators and analysts to search, archive, and analyze speeches as soon as they occur or afterwards. Alongside the AI transcription engine, I also configured and deployed the backend server infrastructure using Sanic, a high-performance Python framework. I designed and implemented all necessary API routes to make the service available to other components of the GIANO platform, ensuring robust communication and a smooth user experience.

The introduction of this speech to text module has significantly improved the efficiency of parliamentary analysis and digital archiving for Nomos. Researchers and users can now quickly access, review, and analyze Senate speeches in real time, opening up new opportunities for data-driven insight, transparency, and innovation. This architecture is also ready for future enhancements, such as speaker identification, sentiment analysis, or advanced search functionalities, further expanding the analytical potential of the GIANO platform in the years to come.