Open source · Runs locally · 34 providers and counting

The dashboard your AI tools forgot to build.

Track spend, quotas, and rate limits across the AI coding tools you actually use. Runs locally, in your terminal.

Get startedDocs
Made for people who run more than one coding agent at a time.

Spend, quotas, rate limits, model usage. One screen.

Runs locally. Keep it beside the tools you already use.

dashboard · detail · compare · analytics views

34 providers, and counting

Coding agents, API platforms, local runtimes. All in one place.

Coding Agents & IDEs
Claude Code
Codex CLI
Cursor
GitHub Copilot
Gemini CLI
OpenCode
Ollama
Amp
Goose
Hermes
Kilo Code
Kimi CLI
Kiro
Roo Code
Zed
Qwen CLI
OpenClaw
Droid
Mux
Pi
API Platforms
OpenAI
Anthropic
OpenRouter
Groq
Mistral AI
DeepSeek
Moonshot
Perplexity
xAI
Z.AI
Google Gemini API
Alibaba Cloud

Yours missing? Open an issue or send a PR.

In your status bar, not another tab

The active tool's usage rides along where you already are — your tmux status bar and the Claude Code statusline. Real provider logos, live cost and quota, nothing to open.

tmux status bar
0:zsh  1:nvim  2:logs5h15%·$6.79/today
Claude Code statusline
Opus 4.8·$0.42 session·$6.79 today·🔥 $1.20/hr·🧠 42%

Set it up in one command: openusage tmux install · openusage statusline --install

Keep it open beside the agent you are using.

OpenUsage next to OpenCode, watching OpenRouter spend live.

What it does

If you only use one coding agent, you don't need this. If you use two or more, you've probably got five tabs open trying to figure out where your money went. That's the problem we solve.

One place across providers

Coding agents, API platforms, and local runtimes side by side. No more cycling through provider dashboards.

Quotas, spend, and limits in one view

Remaining credits, reset windows, rate limits, and model breakdowns where the provider exposes them.

Made for the tools, not the SDK

This is for monitoring the AI tools you use to write code. It's not a tracing SDK for the AI app you're building.

Local by default

No hosted observability plane. The dashboard runs on your machine. History sits in a SQLite file you own.

Beyond the bill

Model breakdowns, session activity, MCP usage, code stats, and daemon-backed history. Where providers expose it.

Open source, growing

A long list of supported providers, more every release. Yours missing? Open an issue or send a PR.

More reading

Comparison

OpenUsage.sh vs OpenUsage.ai

What's the difference. Short answer: different tools, different jobs.

What it tracks

Capability matrix

What it actually tracks, provider by provider. No marketing.

Quick guide

Local quota tracker

The short version, if you found us by searching for "Claude Code quota tracker" or similar.

Docs

Docs and guides

Per-tool guides for Claude Code, Codex, Cursor, OpenRouter, and the rest.

Rearrange dashboard sections. Tune detail graphs. Switch time windows. Set thresholds.

Settings modal. Layout, graphs, thresholds, live preview.

Get started

Install, run openusage, that's it. Auto-detection finds your installed tools and common API key env vars on its own.

Brewbrew install janekbaraniewski/tap/openusage
Scriptcurl -fsSL https://github.com/janekbaraniewski/openusage/releases/latest/download/install.sh | bash
Gogo install github.com/janekbaraniewski/openusage/cmd/openusage@latest

Then run openusage