Antigravity is Google's agentic development platform, launched November 18, 2025 alongside the Gemini 3 model family. PR and marketing teams use Antigravity to build and maintain their own internal tools, trackers and newsroom sites, with an AI agent handling the coding work.
What is Antigravity?
Antigravity is Google's agent-first development platform, built around an editor view for hands-on work and a separate Manager surface where a person spawns, orchestrates and observes multiple coding agents working across projects, according to Google's own developer blog announcement.
Google expanded the platform to Antigravity 2.0 at I/O 2026 in May, adding a standalone desktop application for orchestrating agents and bringing the platform to Google Cloud customers through the Gemini Enterprise Agent Platform, per Google's I/O 2026 developer highlights post. Antigravity is available today at no cost for individuals in public preview.
How does an agent build software without a person writing each line?
An Antigravity agent plans a task, writes the code, uses the terminal to run it, and then uses a browser to test and verify the result, all without a person switching between those tools manually.
Why it works: Antigravity treats the editor, terminal and browser as surfaces a single agent can act across in one session, instead of the agent only suggesting code for a person to run elsewhere, which is the specific mechanism Google describes in its developer platform announcement. The agent reports back through Artifacts, screenshots and walkthroughs of what it built, giving the person reviewing the work enough context to approve it without re-running every step themselves.
Why would a comms team, not an engineering team, use Antigravity?
A comms team without a dedicated engineer would use Antigravity because the same agent that builds a production application can build a much smaller internal tool: a media-monitoring tracker, an approval-routing page, or a static newsroom microsite, at a scope no one would justify hiring a developer for.
Illustrative scenario: a communications team wants a simple internal dashboard that pulls that week's press mentions into one page for a Monday leadership review. Described in plain language to an Antigravity agent, the tool plans the page, writes the code, tests it in a local browser session, and reports back with a screenshot for the team to approve, without a person hand-writing the underlying script.
What is the Manager surface, and why does it matter here?
The Manager surface is Antigravity's dedicated interface for spawning, orchestrating and observing multiple agents working asynchronously across different projects, separate from the Editor view a developer would use for hands-on line-by-line work.
Why it works: The Manager surface flips the usual pattern of an AI assistant embedded inside a single tool, like a chat panel bolted onto an editor, into a standalone command center where a person directs several agents on separate tasks at once and reviews each one's progress independently, according to Google's own description of Antigravity 2.0. For a comms team, that is the specific feature that matters more than the Editor view: a non-technical staffer is far more likely to describe a task to an agent from the Manager surface and check its results later than to sit inside a code editor reviewing individual lines.
Antigravity 2.0, introduced at I/O 2026, made the Manager surface available as its own standalone desktop application separate from the IDE, so a comms staffer working alongside a developer can run Antigravity's agent orchestration without also opening the full code-editing environment the developer uses.
What should a comms team check before going client-facing?
A comms team should check that a technical reviewer has approved an agent's output, that the tool built does not store or transmit client data outside the organization's approved systems, and that the agent's task was scoped narrowly enough to review in full, before letting anything an Antigravity agent built touch client-facing work.
Google's own positioning of Antigravity emphasizes building for user trust across professional developers, hobbyists and everyone in between, with verification tests and Artifacts as the mechanism for that trust, per the platform's product documentation, which signals that the review step is a designed part of the workflow rather than an optional add-on a cautious team invents for itself. A comms department without in-house engineering capacity should still route anything beyond a purely internal prototype through a technical reviewer, whether that reviewer sits on the marketing team, in IT, or at an agency partner.
What does a team need in place before using Antigravity for real work?
A team needs someone comfortable reviewing an agent's Artifacts and verification results before trusting the output, because Antigravity is still built around a review step, not unattended deployment, per Google's own product description of the platform.
Google positions Antigravity for developers, hobbyists vibe-coding in their spare time, and everyone in between, according to the platform's own site, which signals that a non-technical comms staffer can direct it in plain language but still benefits from a technical reviewer checking the agent's work before it touches anything client-facing or handling sensitive data.
Where did Antigravity come from, and is it free?
Antigravity's editor view builds on technology from Windsurf, a coding-assistant startup, and Google unveiled the platform with the return of co-founder Sergey Brin to daily engineering work, according to a KDnuggets technical writeup on the launch. Antigravity remains available at no cost for individuals in public preview as of I/O 2026, with Google introducing a 100 USD per month AI Ultra plan for teams that need to build at a larger scale.
For a comms team, that pricing matters because the entry point is free: a single staffer can open Antigravity, describe an internal tool in plain language, and get a working result back without a procurement conversation, which is a different adoption path than most enterprise software a communications department would otherwise need IT to approve and budget for in advance.
| Internal need | How Antigravity addresses it | What still needs a human |
|---|---|---|
| Media monitoring dashboard | Agent plans, builds and tests the tool across editor, terminal and browser | Review of the Artifact and verification output before rollout |
| Approval-routing internal page | Agent writes and verifies the workflow logic autonomously | Confirming data handling meets the organization's security policy |
| Static newsroom microsite | Agent scaffolds the site and iterates on UI changes from plain-language requests | Final content review and brand sign-off before publishing |
CONCLUSION
Antigravity puts a coding agent in charge of the editor, terminal and browser at once, which is what makes it realistic for a comms team to describe an internal tool in plain language and get a working, tested result back for review, not just a code suggestion to implement themselves. The platform still expects a person reviewing its Artifacts before anything goes live, so it fits internal tooling today, with a technical reviewer in the loop for anything client-facing.
