AI2025
Neura
AI multimedia processing, synthesis & translation platform.
Neura is an AI-powered multimedia processing and transcription platform I build at Nullius In Verba. It turns raw audio and video into transcripts, synthesized speech, subtitles and translated content in real time.
The focus is making AI feel instant and trustworthy — live transcription that keeps up with speech, subtitles that stay in sync, and translation that preserves meaning.
Architecture
- Frontend: Next.js (App Router) + React Server Components, custom hooks for the multimedia player and workflow orchestration.
- AI: custom LLM integrations for text-to-speech (TTS) and speech-to-text (STT), translation and text correction.
- Media: a rendering engine with automatic, synchronized subtitle generation across languages and formats.
- Backend: Supabase (PostgreSQL) for auth, storage, queries and real-time collaboration.
Tech stack
Next.jsReactTypeScriptLLM IntegrationTTSSTTTranslationSupabasePostgreSQL