[Playtime Session] Implement voice input functionality for mic button

[Rushi-faldu]

Purpose

Enable interactive voice-based commands using Genkit’s generative AI framework for real-time user interactions during playtime sessions.

User story

As a parent, I want to use voice commands during a playtime session to repeat instructions, pause, skip, or modify activities, so the interaction feels natural and personalized.

Acceptance criteria:

• Mic button activates voice input using Genkit’s Speech-to-Text (STT) module to capture real-time user queries/commands like Yes, ready, etc for positive confirmation or other commands like 'Repeat', 'Skip', and 'Pause'.
• If the user gives the 'Repeat' command, then the agent fetches the last activity step and reads it aloud.
• If the user gives the 'Pause' command, the agent pauses the current activity and provides a user confirmation message, e.g., "Activity paused. Say 'resume' to continue."
• Map the commands mentioned above with respective actions.
• The agent should also accept personalized modifications from users (e.g., "Can I use blocks instead of cups?") and adapt the instruction dynamically using generative AI and deliver the updated instruction.
• If the voice command is unclear or not recognized, provide a fallback response, e.g., "I didn’t catch that. Could you say it again?".
• Use audio feedback to confirm actions, e.g., "Repeating the instruction." or "Skipping to the next activity."

Other notes

Out of scope:

• If the user gives the 'Skip' command, the agent skips the current activity and prompts the Generative AI agent to suggest the next one.
• If the mic button is not actively used for 5 seconds, it should automatically turn off for privacy and not listen to the user unless the user clicks on it again.

Design notes

• Wireframes are here

• Design specifics are here

---

Definition of Done (DoD)

• The feature has been fully implemented, thoroughly tested, and integrated into the app with no critical bugs or performance issues.
• Manual testing is complete, verifying that the feature meets all acceptance criteria.
• Big features are documented inside the code, with clear explanations of functionality, usage, and any relevant edge cases.
• The feature has undergone peer review, with the code approved by at least one other team member to ensure quality and maintainability.
• The feature branch has been successfully merged into the main branch, and the pull request is closed and the branch is deleted.
• The feature’s UI/UX has been reviewed to ensure consistency and alignment with design guidelines.