Google Antigravity release notes: 2.0.6 links Antigravity 2.0 to the IDE

Google Antigravity has a small but useful changelog entry from May 22: version 2.0.6 adds integration with Antigravity IDE. This is not a big stage announcement. It is more practical than that. It reduces friction when agent work needs to move from the agent environment into a more visual development view.

Google Antigravity is Google's agentic development platform. By agentic IDE, we mean a development environment where AI agents can work with projects, files, terminals and review, while a human still sets scope, permissions, and approvals.

Source: Google Antigravity Changelog

Google Antigravity release notes: what changed in 2.0.6?

Google lists Antigravity 2.0 version 2.0.6 with the date May 22, 2026, and the heading Antigravity IDE integration. The changelog says integration with Antigravity IDE has been added.

The two concrete improvements are:

• A button to install Antigravity IDE when the IDE is not installed.
• A button to open Antigravity IDE when the IDE is installed. Google describes it as a way to open the current project in Antigravity IDE.

Google also says new versions roll out gradually and may take a few days to reach all users. Treat 2.0.6 as a current release-note signal, not something every user necessarily sees at the same time.

Source: Google Antigravity Changelog

Why the IDE link matters for practical teams

This sounds like a button update, but buttons often change habits. If Antigravity is already running agent work inside a project, you want to open that same project in the IDE without losing context. That matters when a developer, technical project lead or automation owner needs to review the work.

A Project in Antigravity is a configuration of folders that defines the agent's environment and scope. Google also says projects have their own isolated settings, and that permissions can be tightened at the project level. That is the right place to think through access, approvals, logs and which commands the agent may run.

For Swedish and Nordic teams, the point is simple: let the agent do the prep work, but make review visible. Once the work opens in the IDE, someone should be able to see changed files, proposed tests and risk points before anything moves forward. In Hammer terms, this belongs in Tool Forge: make the agent workflow useful, but build traceability and human control into it from the start.

Sources: Google Antigravity Changelog and Antigravity Projects documentation

Short example: use the new Antigravity feature

Separate what the human does in the interface from what you ask the agent to do. Google describes the IDE integration as a button: install Antigravity IDE if it is missing, or open the current project in Antigravity IDE if it is already installed.

Human step: Click the button to install or open Antigravity IDE. Use the IDE as the review workspace once the project is open.

Prompt to the agent after the project is open:

Create a short review summary for the latest agent work: which files changed, why they changed, which assumptions should I check manually, and which tests need to run before merge?

Good output should:

• List changed files and why they changed
• Separate facts in the code from the agent's assumptions
• Suggest concrete tests or manual checks
• Mark what needs human approval before merge

Sources: Google Antigravity Changelog and Antigravity IDE getting started

What to watch in the next Antigravity changelog

This release note is about the link between Antigravity 2.0 and the IDE. The next useful signal will be whether Google makes more of the review path feel connected: change summaries, tests, agent logs, permissions, and handoff between terminal, project, and IDE.

If you test Antigravity in a real workflow this week, choose a project where you can measure the difference. How long did review take? Was it clear what the agent did? Did the right person approve the right step? Those answers matter more than another demo.

Source: Google Antigravity Releases