Platform to build functionality for LLM-powered smart home speakers
Tech for EQ (this is what the speaker will use to dynamically modify the EQ based on the audio waveform input):
- Web Audio API - Browser's native audio processing engine (will be swapped to something more robust)
- AudioContext - Main audio processing context
- AudioBuffer - In-memory audio data representation
- AudioBufferSourceNode - Audio playback source
- BiquadFilterNode - Digital audio filters for EQ bands
- GainNode - Volume/gain control
Tech for chatbot (the speaker will run this in the background):
- Web Browser's Speech-To-Text (STT) transcription (will be swapped to something more robust)
- ChatGPT 4o for response generation
- ElevenLabs for Text-To-Speech (TTS) responses (will take ChatGPT responses and apply TTS)
Other tech:
- Supabase for user authentication and database