Google AI Studio

Google's web-based playground for testing Gemini models including Gemini 3.1 Flash Live and Gemini Robotics-ER 1.6. Provides a free, no-API-key environment to build simple voice agents with saved system instructions, voice selection, thinking levels and function calling, and a developer surface for trying embodied reasoning prompts.

Key facts

  • Free tier requires no API key but uses your data for product improvement
  • Switch to 'Live' in the model dropdown to reach Gemini 3.1 Flash Live Preview
  • Supports saving named system instruction presets (persona + style) that can be swapped mid-session
  • Exposes toggles for voice, media resolution, thinking level (off/low/medium/high), session context, function calling and Google search grounding
  • Functions as the lowest-friction starting point before jumping to the Gemini Live API
  • Google DeepMind says Gemini Robotics-ER 1.6 is available in Google AI Studio, with a developer Colab for configuring the model and prompting embodied reasoning tasks [src-039]
  • Google says Google AI Studio now defaults to Interactions API for Gemini model and agent development [src-143].

Related concepts

Source references

  • [src-007] Nate Herk cluster — Nate Herk — Voice AI agents cluster (4 videos)

– Videos referenced: Qt3zMBH-FNg

  • [src-039] Laura Graesser and Peng Xu — "Gemini Robotics-ER 1.6: Powering real-world robotics tasks through enhanced embodied reasoning" (2026-04-14)
  • [src-143] Ali Cevik / Google – "Interactions API: our primary interface for Gemini models and agents" (2026-06-22)

Robin Cartier perspective

This page is part of Robin Cartier's working AI knowledge graph: a practical research layer for production AI, recommendation systems, experimentation, GEO, and agentic web readiness.

The useful next step is to connect this concept back to applied product leadership and operating models.

Recommended next

Keep reading from this thread

From 500 indexed pages and articles.

  1. Wiki concept Agent-First APIs Model-provider interfaces designed for stateful, tool-heavy, long-running agent work rather than single request-response model calls. Related by google
  2. Wiki concept Agentic Vision A visual reasoning pattern where a model uses intermediate actions such as zooming, pointing, and code execution to inspect an image Related by gemini
  3. Wiki concept Physical Safety Constraints for Robots Limits on what a robot should perceive, select, handle, move, or avoid when acting in Related by gemini