What problem does it solve? Adding voice input to Spectacles lenses requires correctly configuring the Lens Studio ASR Module, handling partial versus final transcripts, managing session lifecycles, and avoiding common pitfalls like starting transcription in onAwake. This Skill provides the canonical patterns and runnable components for streaming speech-to-text on Specs. ## Core Features & Use Cases - ASR Session Setup: Configure AsrTranscriptionOptions with silence timeout and accuracy modes (HighAccuracy, Balanced, HighSpeed), then start and stop transcription safely from OnStartEvent. - Transcript Event Handling: Process partial and final transcription updates and handle error codes such as InternalError, Unauthenticated, and NoInternet. - Voice-to-AI Pipelines: Route final transcripts into Gemini or OpenAI assistants for chat, query, and summarization experiences. - Use Case: Build a voice-controlled chat lens where a user speaks a question, the ASR Module finalizes the transcript after 1-2 seconds of silence, and the text is sent to an LLM for a spoken or displayed response. ## Quick Start Use the specs-asr skill to add real-time speech-to-text transcription to my Spectacles lens component.
