Building a cooking assistant with Firebase AI Logic

A Friendly Meals Android app prototype demonstrates how Firebase AI Logic and the Gemini Live API can create a hands-free cooking assistant that processes live video and spoken questions, responds through streamed audio, and uses the current recipe as context to assess cooking progress. Firebase AI Logic provides secure client SDKs and manages bidirectional streaming without requiring a custom backend, while Android camera and microphone permissions enable the live interaction. The implementation configures a LiveGenerativeModel with an audio voice, recipe-specific system instructions, and model settings stored in Firebase Remote Config so they can be updated without an app release. Video frames are compressed and sent at a limited rate to conserve bandwidth, while client-side function calling allows the model to identify requests such as adding garlic or olive oil to a grocery list and trigger an in-app database update for the signed-in user. The post also recommends Firebase App Check for production protection and notes that prompt configurations should use Server Prompt Templates where supported.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.