ai-remote-service

Integrate AI APIs for chat and voice in Spectacles by loading the RemoteServiceGateway.lspkg package and establishing a WebSocket-based session. This unifies Gemini Live and OpenAI Realtime workflows without embedding API keys.

Yes, you can use OpenAI Realtime function calling without embedding API keys by orchestrating AI tools through the Remote Service Gateway. It handles loading, authenticating, and routing function declarations securely.

The best way to set up WebSocket sessions for Gemini Live voice and vision is to load the RemoteServiceGateway.lspkg package and initialize a session following the provided reference implementations for Spectacles projects.

Yes, the Remote Service Gateway supports image generation workflows alongside chat, voice, vision, and function calling. It provides integration patterns for DALL-E and Snap3D within Spectacles projects.

You need a WebSocket-based session for OpenAI Realtime integration to maintain a persistent, low-latency connection required for real-time voice interactions and immediate function calling responses.

When orchestrating multiple AI tools with RSG, you must manage WebSocket session states carefully and ensure the RemoteServiceGateway.lspkg package is properly loaded to avoid authentication and connection drops across concurrent workflows.