MOMO_AI

AI & systems / Sep 2025

MOMO_AI

A collaborative multimodal assistant connecting wake-word detection, speech recognition, language models, and voice output.

My contribution

Contributed AI/ML components, data-processing experiments, debugging, and architectural refinement in a collaborative fork.

The approach

• Detect a wake word, capture speech, and transcribe it with faster-whisper.
• Route requests through Gemini and developer-command modules.
• Return spoken responses using ElevenLabs, with desktop UI and session handling.

Scope & perspective

Voice interaction ties together audio timing, transcription, command handling, and feedback. The expressive 3D avatar remains an extension rather than a completed integrated feature.

Collaborative project; this fork documents Arnav's contributions and experiments.