OnWatch – Track 6 AI API quotas from your terminal (<50MB RAM, zero telemetry)

2 points

3 months ago

I built onWatch because I was checking 6 different AI provider dashboards every day to track my quota usage. Each has different billing cycles, different formats, and none show historical patterns.

onWatch is a Go CLI that runs as a background daemon, polls your API quotas (Anthropic, OpenAI Codex, GitHub Copilot, Synthetic, Z.ai, Antigravity), stores history in SQLite, and serves a Material Design 3 web dashboard.

Key decisions:

- Single binary (~13MB), no runtime dependencies - <50MB RAM with all 6 providers polling in parallel - All data stays local — zero telemetry, no cloud - One-line install for macOS, Linux, Windows - Docker support (distroless, non-root, ~10MB image)

The insight I kept coming back to: knowing your current usage isn't enough. You need historical cycle data — which sessions burn quota fastest, how usage compares across billing periods, and when to expect resets.

Install: `curl -fsSL https://raw.githubusercontent.com/onllm-dev/onwatch/main/install.sh | bash`

https://github.com/onllm-dev/onwatch