# ng-firebase-tts

> A client-side Angular application demonstrating real-time streaming text-to-speech (TTS) and multimodal image analysis powered directly by Firebase AI Logic (Vertex AI for Firebase) and Gemini models with zero server-side orchestration.

## Core Documentation

- [Full Documentation](/llms-full.txt): Complete in-depth architectural and technical specification for AI agents.
- [Interactive Application](/): Web application interface for multimodal vision analysis and real-time text-to-speech streaming.

## Features & Modules

- [Real-Time Speech Synthesis](/): Zero-latency streaming text-to-speech with Gemini Voice Models (`Aoede`, `Charon`, `Fenrir`, `Kore`, `Puck`, `Zephyr`) and Web Audio API gapless scheduling.
- [Multimodal Image Comprehension](/): Automated extraction of descriptive tags, accessibility alt-text, image suggestions, and Google Search grounded trivia.
- [Vocal Customization](/): Contextual parameters for Scene Description, Vocal Emotion, and Speaking Pace.
- [PWA & Offline Caching](/manifest.webmanifest): Service Worker caching manifest for full offline shell capability.

## Optional & Configuration

- [Firebase SDK Config](/firebase.config.json): Client-side Firebase initialization configuration.
- [Remote Config Defaults](/remote-config-defaults.json): Default parameters for Gemini models, thinking levels, and locations.