Vocalize — AI Speech Processing Platform

Vocalize is an end-to-end full-stack web application designed for high-precision audio ingestion, speech-to-text (STT) neural transcription, interactive word-level audio-text synchronization, and multi-format subtitle exports. Built with a clean light-mode dashboard interface, Framer Motion UI animations, and automated multi-provider AI adapters.

Key Features

• Multi-Provider AI STT Adapters: Automatically uses Google Gemini 3.6 Flash Multimodal Audio API, Groq Cloud Whisper-Large-v3, or OpenAI Whisper-1 with deterministic mock fallbacks.
• Word-Level Audio & Text Synchronization: Every word in the transcript is timestamped to the millisecond. Clicking any word in the transcript jumps audio playback directly to that exact millisecond.
• Live Active Word Highlighting: Audio playback highlights spoken words in real time with high-visibility sage-green accents.
• Interactive Audio Waveform Visualizer: Real-time canvas waveform scrubber with play/pause controls, speed selectors (0.5x, 1x, 1.5x, 2x), and optional AI Voice synthesis.
• Multi-Format Subtitle Exporter: Instant client-side download to Plain Text (.txt), Structured JSON (.json), SubRip Subtitles (.srt), and WebVTT Subtitles (.vtt).

Technical Details

• Frontend: React 19, TypeScript 5.x, Vite 6, Tailwind CSS v4, Framer Motion 12, Lucide Icons.
• Backend: Node.js 20+, Express.js REST API, Multer audio stream buffer guard, X-Request-ID stamping.
• Containerization & Cloud: Multi-stage Docker containerization, GCP Cloud Run continuous deployment with GitHub Actions.