🚨 Critical Regression in Latest Antigravity Version – Chat Freeze, Conversation History Loss & Pro Plan Limit Concerns (Windows 10)

Google Antigravity
, , , ,
1

Environment

OS: Windows 10 (22H2)
Antigravity Version: Latest Release ON 21 FEB 2026

Summary

After updating to the latest Antigravity version, I am experiencing severe performance degradation in long chat sessions, disappearing conversation history from the UI, session persistence issues, and restrictive Claude model limits under the Pro subscription tier.

These issues significantly impact workflow continuity and overall product trust.

:one: Severe UI Lag in Long Chats

Observation:

  • As conversation length increases, the UI becomes progressively slower.

  • Large sessions freeze or become unresponsive.

  • Performance degradation appears proportional to chat size.

  • Previous versions handled long chats noticeably better.

Concern:

The system may be rescanning full conversation history instead of using incremental context handling.

Suggestion:

  • Implement incremental context summarization.

  • Introduce sliding window context management.

  • Lazy-load older messages in UI.

  • Move indexing to background async processing.

This would significantly improve scalability for long-term sessions.

:two: Conversation History Disappearing from UI

Observation:

  • Several past conversations are no longer visible in the UI.

  • Corresponding conversation folders still exist locally.

  • Restarting the application does not restore them.

  • Restarting multiple times in the same day increases the issue likelihood.

Likely Cause:

  • Index corruption.

  • UI-layer desynchronization from local storage.

  • Session state mismanagement.

This creates a high-trust issue, as users may believe critical data has been permanently lost.

:three: Session Persistence Instability

  • Restarting Antigravity multiple times within the same day worsens chat visibility issues.

  • Suggests improper cleanup or state handling between sessions.

  • Possibly memory leak or indexing inconsistency introduced in latest version.

:four: Clear Regression from Previous Versions

  • Earlier versions were significantly more stable.

  • Long sessions were usable.

  • Conversation history was reliably visible.

The latest update appears to introduce architectural instability.

:five: Pro Plan Claude Model Session Limits

Observation:

  • Google AI Pro subscribers receive only 2 sessions on Claude models.

  • After consuming these sessions, access is blocked for approximately 5–6 days.

  • Session recycling appears significantly slower than at initial launch.

Concern:

The current limits feel restrictive relative to expectations from a “Pro” tier subscription.

Impact:

  • Limits meaningful development usage.

  • Interrupts workflow continuity.

  • Reduces perceived value of subscription tier.

Request:

  • Transparent documentation of:

    • Exact session limits

    • Reset timing

    • Model-specific quotas

  • Consider daily refresh or token-based quota instead of long cooldown periods.

  • Provide visible quota dashboard for users.

Overall Impact

  • Workflow disruption

  • Forced creation of new chats

  • Reduced long-session usability

  • Data trust concerns

  • Subscription value inconsistency

These issues significantly affect power users relying on Antigravity for continuous development workflows.

Thank you for reviewing this. I hope these observations help improve stability and long-session scalability in future updates.

2

Hello @ANURAJ_RAI , welcome to AI Forum!
Thank you for sharing your observation with all the details. We have escalated the issue to our internal teams for a thorough investigation.

To ensure our engineering team can investigate and resolve these issues effectively, we highly recommend filing bug reports directly through the Antigravity in-app feedback tool. You can do this by navigating to the top-right corner of the interface, clicking the Feedback icon, and selecting Report Issue.

1 Like