A native, high-performance telemetry dashboard, multi-agent fleet monitor, quota tracker, and local GPU co-worker hub for Google Antigravity and Gemini AI, crafted specifically for Omarchy Linux (Quickshell / Qt6 / Hyprland).
Authors: simonez & Arci
Version: 1.6.0
License: MIT
| ⚡ 1. Performance & Quotas | 📂 2. Sessions & Fleet | ⚙️ 3. Settings & Telemetry |
|---|---|---|
![]() |
![]() |
![]() |
- Minimalist Top Bar Slot:
- Clean status badge matching native Omarchy bar widgets (
10pxtypography, compact footprint). - Live heartbeat pulse animation active exclusively while the AI agent is thinking/generating code, with instantaneous idle reset upon turn completion.
- Clean status badge matching native Omarchy bar widgets (
- Multi-Account Auto Failover (Tri-Pool):
- Automatic seamless failover between 3 Google AI Pro accounts (
agy,agy2,agy3) upon individual quota exhaustion. - Instant account state synchronization and interactive account switcher in both the dashboard and launcher.
- Automatic seamless failover between 3 Google AI Pro accounts (
- Bounded I/O & Memory Ceilings:
- Strict memory ceilings (
2 MB / 512 KB / 64 KB) preventing buffer overruns. - Path traversal protection (
.is_relative_to()) and leak-free socket connections for GCP latency pings. - Calibrated subpixel sparkline rendering and layout geometry.
- Strict memory ceilings (
- Google AI Pro Quota Gauges: Real-time 5-Hour Session Quota (2.5M CAP) and 7-Day Weekly Quota (25M CAP) with live Free Headroom gauges and reset countdowns.
- 1M Context Window Breakdown: Stacked horizontal meter visualizing System & Rules, Tool & MCP Schemas, File & Code Context, Conversation History, and Free Headroom.
- 7-Day Sparkline Trend: Custom Canvas trend curve with translucent gradient fill, continuous polyline, and glowing data points.
- Google Cloud Edge Latency: Live connection checks and ping latency (~14 ms) for Gemini Interactions API, Grounding/Search, Codebase Embeddings, and Agent Fleet.
- Recent Sessions Manager: History of recent conversation sessions with 1-click Copy ID and instant Resume/Open launcher.
- Local GPU Co-Workers: Zero-token local inference co-worker monitoring on NVIDIA RTX 3070 CUDA (
coder,auditor,bonsai) and Groq Cloud fast inference. - Subagents Fleet Profiler: Real-time metrics for autonomous subagents (
sec-auditor,qml-designer-reviewer,test-runner,doc-researcher) with latency badges, speed scores, and cumulative cloud tokens saved. - Stacked Tool Calls Breakdown: Granular distribution meter of real-world tool executions (
view_file,run_command,replace_file_content,grep_search). - Beads Rust Integration: Dedicated quick-action button launching the interactive TUI Kanban and DAG project graph (
bv).
- Auto-refresh Presets: Configurable scan frequency (
10s,30s,60s,120s,300s). - Working Pulse Toggle & BPM: Heartbeat animation toggle and customizable cadence presets (Sleep 40, Rest 60, Walk 85, Sprint 130 BPM).
- Desktop Notifications: Native Wayland notifications (
notify-send) upon task completion.
cd ~/Projects/Antigravity1.1
./install.shOr via Omarchy Plugin Marketplace:
omarchy plugin add https://github.com/simonezpx3/antigem.git --enablecd ~/Projects/Antigravity1.1
./uninstall.shThe plugin includes a dedicated Wayland launcher script for instant session recovery and account dispatching:
omarchy-launch-antigravity # Launch interactive Antigravity session
omarchy-launch-antigravity resume <id> # Resume specific conversation by session ID
omarchy-launch-antigravity --status # Display current quotas and active model

