minute with Mistral, who are sponsoring this video. First, let's prototype this in Mistral Studio. In the agent playground, I'll add it two mock functions, one to start the timer and another one to stop it. Then I'll test out whether the Mistral LLM knows to call them at the right time based solely on their description. It looks like it can, so we're ready to code this up. I'll head over to Mistral Vibe and ask it to implement a cascaded voice assistant that chains pre models Vox Troll real time for speech to text Minstrel 3B as an LLN and Vox Troll DTS for text to speech. And of course the timer function that I just tried out. And in just 108 seconds we built a voice