Question 1

What is OpenUsage?

Accepted Answer

OpenUsage is a free, open-source LLM dashboard that tracks AI agent costs and token usage from your terminal. It auto-detects 17 installed tools like Claude Code, Cursor, Copilot, Codex CLI, Gemini CLI, and OpenRouter, then displays live cost tracking, quotas, model burn rates, and token consumption in a keyboard-driven terminal interface.

Question 2

How does OpenUsage track LLM costs?

Accepted Answer

OpenUsage provides LLM cost tracking by auto-detecting your AI coding tools and API keys, then polling provider APIs for real-time spend data, rate limits, and quota usage. It tracks agent costs across Claude Code, Cursor, Copilot, Codex CLI, and 13 other providers. The daemon mode continuously monitors token consumption and stores cost history in a local SQLite database for trend analysis.

Question 3

Can OpenUsage track AI agent token usage?

Accepted Answer

Yes. OpenUsage tracks agent token usage and consumption per model, per provider, and per session. It shows token distribution breakdowns, input vs output token counts, cost per token, and burn rates. For tools with local telemetry (Claude Code, Cursor, Codex, Gemini CLI, Copilot, Ollama), it captures detailed per-session token consumption data.

Question 4

Which AI coding tools does OpenUsage support?

Accepted Answer

OpenUsage supports 17 providers across three categories for LLM cost tracking. Coding agents: Claude Code, Codex CLI, Cursor, Copilot, Gemini CLI, OpenCode, and Ollama. API platforms: OpenAI, Anthropic, OpenRouter, Groq, Mistral, DeepSeek, xAI, Z.AI, Gemini API, and Alibaba Cloud.

Question 5

How do I install OpenUsage?

Accepted Answer

Install via Homebrew (brew install janekbaraniewski/tap/openusage), install script (curl -fsSL from GitHub releases), or Go install (go install github.com/janekbaraniewski/openusage/cmd/openusage@latest). First run auto-detects all installed providers with zero configuration. Just run openusage to start the LLM dashboard.

Question 6

Is OpenUsage free and open source?

Accepted Answer

Yes. OpenUsage is fully open source and free to use for LLM cost tracking and agent usage monitoring. The source code is available on GitHub at github.com/janekbaraniewski/openusage. It is written in Go and uses Bubble Tea for the terminal UI.

Question 7

What is OpenUsage daemon mode?

Accepted Answer

Daemon mode runs a background process that continuously tracks AI agent costs and token consumption, even when the dashboard is closed. It stores telemetry in a local SQLite database with 30-day retention and supports hook integrations for Claude Code, Codex, and OpenCode to capture session-level token usage data.

Question 8

What data does the OpenUsage LLM dashboard track?

Accepted Answer

The OpenUsage LLM dashboard tracks agent costs, spend limits, quota usage, model burn rates, token consumption and distribution by model, tool and MCP usage, code statistics (lines added/removed, languages used), daily cost trends, and session-level telemetry. All data stays local on your machine.

Question 9

Does OpenUsage track MCP tool usage?

Accepted Answer

Yes. OpenUsage tracks MCP (Model Context Protocol) tool usage per session for providers that support it. You can see which MCP tools were called, how often, and in which sessions. MCP usage tracking is available for Claude Code, Codex CLI, and OpenCode through their hook integrations.

Question 10

Can OpenUsage monitor model usage across multiple providers?

Accepted Answer

Yes. OpenUsage provides model usage tracking across all 17 supported providers. It shows which models you're using (GPT-4o, Claude Sonnet, Gemini Pro, etc.), token consumption per model, cost per model, and burn rates. The analytics view lets you compare model usage patterns across providers and time windows.

Question 11

Does OpenUsage track API rate limits and quotas?

Accepted Answer

Yes. OpenUsage provides quota monitoring and rate limit tracking for API-based providers. It reads rate limit headers from OpenAI, Anthropic, Groq, Mistral, DeepSeek, xAI, and Gemini API to show remaining requests, token limits, and quota usage. For subscription-based tools like Cursor and Copilot, it tracks plan limits and usage percentages.

Question 12

How does OpenUsage compare to other LLM observability tools?

Accepted Answer

OpenUsage is a lightweight, terminal-native LLM observability tool focused on local AI coding agents. Unlike cloud-based LLM monitoring platforms, OpenUsage runs entirely on your machine with zero external dependencies. It provides AI agent monitoring, API cost tracking, model usage analysis, and session tracking without sending any data to external servers.

Track AI coding usage, spend, and quotas from your terminal.

What it shows

Spend, credits, and quotas

Token and model breakdowns

Session and client activity

MCP and tool usage

Daemon-backed history

Code statistics

17 providers

Get started