Real-time status bar for the OpenCode sidebar footer: cache hit rate, token usage, cost, latency, and model info — refreshed live via message.updated events (no polling).
- Cache hit rate for the last request and the whole session
- Token usage: input, output, reasoning (per request and cumulative)
- Cost per request and cumulative session cost
- Latency of the last response
- Model name of the last request
- 3-line compact layout, color-coded via your active theme
- Event-driven, zero-polling real-time updates
opencode plugin opencode-status-trackerOpenCode downloads the package from npm, registers it in tui.json, and is
ready after restart.
Add the plugin to your tui.json (OpenCode auto-installs it from npm at startup):
{
"$schema": "https://opencode.ai/config/tui.json",
"plugin": ["opencode-status-tracker"]
}{
"plugin": ["./opencode-status-tracker/src/tui.tsx"]
}Restart OpenCode and open the sidebar (Ctrl+X B) to see the footer.
✓ deepseek-v4-flash · 4.2s · $0.003
last: in 12.0k · out 1.2k · hit 86% · rd 10.0k · wr 2.0k
📊 total: 8 req · in 70.0k · out 9.0k · hit 78% · $0.28
hit= cache hit rate =cache.read / (input + cache.read)rd/wr= cache read / write tokens- Token units auto-scale:
k/m/b
npm install
npm run build # esbuild -> dist/
npm run typecheck # tsc --noEmit- OpenCode >= 1.4.3 (sidebar footer slot)
- Peer deps:
@opencode-ai/plugin,@opentui/core,@opentui/solid,solid-js
MIT