I want to bring a casual, friendly AI host to my online radio station—one that grabs the latest local and national headlines (plus bits from nearby high schools), turns them into lifelike speech on par with ElevenLabs, and drops the finished audio straight into my stream. Between songs the host should also handle quick weather and sports updates and slip in a few musical recommendations so the station always feels live and up to date. Because I’m new to AI, the first step is a short discovery chat where you explain what’s realistically possible, outline the tech stack you’d suggest—whether that’s GPT-style summarisation, ElevenLabs or similar TTS APIs, Python or Node for automation, and integration with my existing stream software—and map out ongoing hosting costs. Once we agree on the approach, I’ll need: • A script or micro-service that fetches the chosen feeds, cleans and summarises the text, generates high-quality MP3s, and queues them for broadcast. • A simple way for me to tweak voice settings, update source feeds, and schedule how often new segments air. • Clear documentation so I can run and maintain everything without deep coding knowledge. Acceptance criteria: the voice must sound naturally casual and friendly, headline summaries must be fact-checked, and every segment needs correct metadata and reliable delivery to the station. If you’ve built something similar or know the space well, tell me how you’d tackle this headline-to-air pipeline—I’m ready to get started.