Google releases Managed Agents for the Gemini API, introducing the Antigravity agent on Gemini 3.5 Flash that runs in isolated Linux sandboxes with code execution, file management, and unrestricted web access
Preview offers free compute on pay-as-you-go pricing after Deep Research.
More integrations between Antigravity and Gemini API for easy managed agents!
AIS: https://aistudio.google.com/prompts/new_chat?model=antigravity-preview-05-2026 API Quickstart: https://ai.google.dev/gemini-api/docs/managed-agents-quickstart
Antigravity Agent and Gemini Managed Agents are now rolled out to all users in @GoogleAIStudio and Gemini API. Links below⤵️
Antigravity Agent and Gemini Managed Agents are now rolled out to all users in @GoogleAIStudio and Gemini API.
Links below⤵️

1/ Antigravity Agent. General-purpose managed agent powered by Gemini 3.5 Flash. Runs code, searches the web, and reads/writes files in Google-hosted sandboxes.

I'm excited to introduce Managed Agents in the Gemini API. One API call gives you a full agent with code execution, web browsing, and file management in an isolated sandbox. - Powered by Gemini 3.5 Flash and Google's Antigravity harness - Runs Bash, Python, and Node.js in isolated sandboxes - Define custom agents with AGENTS.md and SKILL.md files - Mount GitHub repos, GCS buckets, or inline files into environments
I'm excited to introduce Managed Agents in the Gemini API. One API call gives you a full agent with code execution, web browsing, and file management in an isolated sandbox.
- Powered by Gemini 3.5 Flash and Google's Antigravity harness - Runs Bash, Python, and Node.js in isolated sandboxes - Define custom agents with AGENTS.md and SKILL.md files - Mount GitHub repos, GCS buckets, or inline files into environments

2/ Environments persist across turns. First call provisions the sandbox. Pass the `environment_id` back and the agent resumes with all files, packages, and state intact.

1/ Antigravity Agent. General-purpose managed agent powered by Gemini 3.5 Flash. Runs code, searches the web, and reads/writes files in Google-hosted sandboxes.
4/ Create custom agents by ID. Declare instructions and skills inline, or fork an active environment into a reusable agent.

3/ Skip the setup loop. Mount GitHub repos, GCS buckets, or inline files directly into the sandbox. The environment arrives ready to use.
3/ Skip the setup loop. Mount GitHub repos, GCS buckets, or inline files directly into the sandbox. The environment arrives ready to use.

2/ Environments persist across turns. First call provisions the sandbox. Pass the `environment_id` back and the agent resumes with all files, packages, and state intact.
Available in preview now (Rolling out at the moment, should be available for everyone ~2:30pm).
→ Blog: https://blog.google/innovation-and-ai/technology/developers-tools/managed-agents-gemini-api/ → Docs: http://ai.google.dev/gemini-api/docs/agents
5/ Network allowlists and credential injection. Restrict which domains the sandbox can reach. Auth headers get injected by the egress proxy, never exposed inside the sandbox.
5/ Network allowlists and credential injection. Restrict which domains the sandbox can reach. Auth headers get injected by the egress proxy, never exposed inside the sandbox.

4/ Create custom agents by ID. Declare instructions and skills inline, or fork an active environment into a reusable agent.
The official @OfficialLoganK getting the #googleio dev keynote kicked off announcement Managed Agents in the Gemini API.

Ugh wrong photo!

The official @OfficialLoganK getting the #googleio dev keynote kicked off announcement Managed Agents in the Gemini API.