Google Antigravity TOS: 3rd party usage can get Google account suspended

Original Article ↗ Hacker News Discussion ↗

Google’s new terms for its Antigravity AI coding tools warn that using third‑party clients can lead to account suspension, alarming developers who fear losing access to critical services tied to their Google accounts such as Gmail, Drive, and YouTube. Although a Antigravity team member later claimed suspensions apply only to the Antigravity service itself, many see the wording and Google’s history of opaque, hard‑to‑appeal bans as too risky. Commenters increasingly advocate de‑risking from Google—by using alternative AI providers, self‑hosting or moving email to independent services, and calling for regulation or structural limits on Big Tech’s power over digital identity.

Scope of the ToS and Account Bans

  • Core controversy: Antigravity’s ToS says use via third‑party tools or non‑Google products can lead to account suspension.
  • Many interpret this as risking their entire Google account (Gmail, Photos, etc.), which they consider too valuable.
  • An Antigravity team member publicly claimed the suspension applies only to the Antigravity account, not the whole Google account, and said the wording will be clarified.
  • Some commenters argue that Google’s history of account‑wide bans makes this reassurance hard to trust.

Third‑Party Tools, CLIs, and Harnesses

  • Debate over what counts as “third party”: personal harnesses, self‑hosted frontends, or using OAuth tokens outside official clients.
  • Some say automation over the official agy CLI is explicitly documented and allowed; others report that such usage has been detected and led to suspensions, especially when using bundled “Pro” quota.
  • Confusion around IDE extensions: official Antigravity extensions are clearly allowed; unofficial VS Code extensions and tools like Pi or T3Code and how they authenticate are seen as risky.
  • Several developers say they’ve removed Antigravity from their orchestrations or will only use it on burner accounts.

User Hostility and Centralized Risk

  • Many see banning the entire Google account for AI ToS violations (even if only possible) as disproportionate and user‑hostile.
  • Repeated anecdotes of Google accounts banned or locked with opaque or impossible recovery processes, sometimes for child‑safety scans or “suspicious” behavior.
  • This risk is heightened because one Google account often controls email, identity, YouTube, Android, payments, and even government or school logins.

Alternatives, Workarounds, and Policy Ideas

  • Multiple commenters refuse to use Google AI at all, or cancel paid AI subscriptions, preferring OpenAI, Anthropic, or open‑weight models.
  • Advice trends: de‑Google where possible, use custom domains with independent email providers, keep regular Google Takeout backups, or isolate AI usage to separate accounts.
  • Broader discussion about regulation: calls for service‑specific bans only, due process, treating email/identity as utilities, and limiting big platforms’ power over digital identity.