Repolode
Submit
JingbiaoMei

JingbiaoMei/Tokdash

Local dashboard that tracks token usage, session costs and subscription quota windows across AI coding agents such as Claude Code, Codex and OpenCode.

View on GitHubWebsite
Stars
92
Forks
21
Open issues
8
Last commit
2 days ago

What is it?

Tokdash is a local token and cost dashboard for AI coding tools. It reads the session logs that clients such as Claude Code, Codex, OpenCode and Gemini CLI write on your own machine, then shows exact input, cache and output token breakdowns, per-session drill-downs, and estimated costs.

The same local data feeds four views: a FastAPI-served WebUI with heatmaps and week/month/year reports, a terminal dashboard launched with `tokdash tui`, a native companion app for macOS menu bar and Windows notification area, and a statusline item that can sit inside an agent's prompt bar. Costs come from a bundled pricing database that can be overridden locally.

Quota tracking is optional. Codex's rolling windows are estimated from local logs, while other providers' subscription windows require opt-in live polling using the sign-in your CLI already has. Installation is via pipx with a `tokdash setup` command that registers a local background service.

Key features

  • 01Exact input, cache and output token breakdowns per session, tool and model.
  • 02Contribution heatmap and week, month or year usage reports exportable as PNG.
  • 03Quota tab showing subscription window bars and reset countdowns for multiple providers.
  • 04Terminal dashboard with seven panes, including Overview, Sessions, Quota and Pricing.
  • 05Statusline integration and a native macOS and Windows companion app.
  • 06Can combine usage from several Tokdash instances running on different machines.

Use cases

  • Checking how much a team's AI coding agents spend each day or month.
  • Comparing token usage across Claude Code, Codex and other agents on one screen.
  • Keeping an eye on subscription quota resets without opening a browser.
  • Keeping local agent history from being deleted by client-side retention settings.

Quick start

  1. 1

    Install the package with pipx

    pipx install tokdash
  2. 2

    Run setup to register the local service and get the dashboard URL

    tokdash setup

Things to know

  • Quota tracking is optional and needs extra setup for each provider.
  • Requires Python 3.10 or newer.

In collections

Claude Code tools, skills & pluginsAI coding assistants & agentsCLI & terminal toolsDashboards & business intelligence toolsOpenTelemetry & observability platforms

Topics

#token-usage#ai-coding#dashboard#cost-tracking#claude-code#codex#observability#cli#agent#analytics

FAQ

What is Tokdash?
Local dashboard that tracks token usage, session costs and subscription quota windows across AI coding agents such as Claude Code, Codex and OpenCode. Tokdash is a local token and cost dashboard for AI coding tools. It reads the session logs that clients such as Claude Code, Codex, OpenCode and Gemini CLI write on your own machine, then shows exact input, cache and output token breakdowns, per-session drill-downs, and estimated costs.
Is Tokdash free to use? What license does it have?
Tokdash is open source under the MIT license. Check the license terms before using it in commercial products.
Is Tokdash still maintained?
Actively maintained. The last commit to Tokdash was on Nov 6, 2026.
What can I use Tokdash for?
Checking how much a team's AI coding agents spend each day or month. Comparing token usage across Claude Code, Codex and other agents on one screen. Keeping an eye on subscription quota resets without opening a browser. Keeping local agent history from being deleted by client-side retention settings.
How do I install Tokdash?
Install the package with pipx: pipx install tokdash Run setup to register the local service and get the dashboard URL: tokdash setup
Where can I get Tokdash?
The source code and installation instructions are on GitHub: https://github.com/JingbiaoMei/Tokdash

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects