Story Spinner is a modular AI-driven storytelling device built on a Raspberry Pi Zero 2 W with a custom enclosure and dedicated display controller PCB. The system generates dynamic stories, character artwork, and expressive narration in real time using the Gemini API. The text-to-speech pipeline is built around token-aware streaming rather than static sentence synthesis. As the language model generates tokens, they are streamed directly into the voice engine, allowing it to interpret punctuation, phrasing, and narrative structure in context. This enables natural inflection, pacing, and emotional continuity across sentences instead of robotic, fragmented speech. By separating interface logic, generative text processing, and audio synthesis layers, the architecture maintains real-time responsiveness while preserving story-level context and expressive delivery.