Overview
Through the codelab, you will employ a step-by-step approach as follows: → → Utilize an existing Hugging Face Dataset of Yoga poses (JSON format). → Enhance the dataset with descriptions generated by the Gemini API. → Use Langchain and Firestore integration to create a collection with vector embeddings in Firestore. → Create a composite index in Firestore for efficient vector search. → Build an interactive Flask web application featuring: → Vector search for pose recommendations with metadata filtering (e.g., expertise level). → Real-time audio instruction generation using the Gemini Live API with voice selection. → Conversational follow-up capabilities for audio instructions. → Web search integration using the Gemini Live API and Google Search tool for broader yoga queries with audio responses. → Text-to-image generation related to web search conversations. → Deploy the application (optionally) to Google Cloud Run.
Join Stepwik
Build and share what you know
Create interactive labs and courses that help developers learn faster — then publish them to your network.
- Create labs
- Share with learners
- Grow your audience
