Tokdash CLI
Tokdash shows token usage and session cost for Claude Code, Codex, Gemini CLI and a dozen more agents locally
- 92 ★ on the repository
- Last commit
Extended access
Why this level
- Tokdash reads agent session logs, registers a background service and, with consent, calls provider quota APIs using the sign-in saved by the CLI.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/tokdashDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Tokdash reads agent session logs, registers a background service and, with consent, calls provider quota APIs using the sign-in saved by the CLI.
This is third-party code. Review the repository files before installing.
What it does
Tokdash is a local token and cost dashboard for coding agents. It reads session logs of Claude Code, Codex, Gemini CLI, OpenCode, Cline, GitHub Copilot CLI and other clients straight on your machine and builds a single index. The same data appears in four views: a FastAPI web dashboard with an overview, sessions, stats, a weekly and monthly report, quotas and a servers tab; the `tokdash tui` terminal dashboard; a menu bar app on macOS and a notification area app on Windows; and a statusline item inside the agent itself. Tokdash shows subscription quotas for nine providers, including Claude Code, Kimi Code, Grok and Z.ai. Codex works from local logs, the rest need optional polling: Tokdash calls the provider quota endpoint with the sign-in the CLI already has, enabled with per-provider consent. Several machines combine into one dashboard, for example WSL and a Mac. Cost is computed from a bundled price database and may lag real tariffs.
Who it is for. Developers who pay for Claude Code, Codex and other subscriptions and want to see token use, cost and remaining quota per agent.
Good fit when
- A subscription limit drains faster than expected and you need to find which session or model used the most
- You want today's spend and the quota window reset timer visible in the menu bar
- You work with several agents at once and want to compare their usage in one dashboard
- Work machines are spread across WSL, a Mac and a server and you need one combined total
- An agent should show token use in its own statusline: the repository has ready bash and PowerShell templates
Not a fit when
- You need team cost tracking on a server with shared billing: Tokdash reads local logs of a single user
- You need an exact figure for accounting: prices come from a bundled database and may lag real ones
- You work on native Windows without WSL and need stability: native Windows support is still experimental
Example request
Install Tokdash with pipx, run tokdash setup and show the dashboard address so I can see Claude Code session usage for the week.Limitations
Python 3.10 or newer is required. Linux, WSL2 and macOS are supported while native Windows is experimental. Data comes from local client logs, so Claude Code and Gemini CLI delete sessions older than about 30 days by default: the author suggests setting `cleanupPeriodDays` to 3650 in `~/.claude/settings.json`. Cost is computed from a bundled price database and may lag provider tariffs. Quota polling is off by default and uses the sign-in already stored in each provider CLI, so enable it deliberately. `tokdash setup` registers a user-level background service and listens on 127.0.0.1, port 55423.
How to disable. Run `tokdash uninstall`: it reverses everything `tokdash setup` created and keeps usage history by default. The `--purge` flag deletes the history too.
FAQ
How do I install Tokdash?
Run `pipx install tokdash`, then `tokdash setup`. The wizard detects the environment, installs a managed runtime if needed, registers a local service and prints the dashboard address `http://127.0.0.1:55423`.
Which agents are supported?
Among them OpenCode, Codex, Claude Code, Gemini CLI, Cline, GitHub Copilot CLI, Kilo Code, Hermes, Kimi CLI and Grok Build. The full list is in `docs/reference/SUPPORTED_CLIENTS.md`.
Does Tokdash send my logs anywhere?
The README says the index is stored locally and the dashboard listens on 127.0.0.1. Data leaves the machine only during provider quota polling, which is off by default.
How do I see usage in the terminal?
`tokdash tui` opens seven panes: overview, sessions, stats, report, speed, quota and pricing.
How do I open the dashboard from another device?
Run `tokdash setup --bind 0.0.0.0`: the dashboard opens on all network interfaces with writes disabled. Read the remote access guide first.
Discussion
No questions or reviews yet.
Related
CLI toolsCounts token use and cost of Claude Code and Codex sessions from local logs and shows remaining limits in the status line
htop for AI costs: a terminal dashboard shows tokens and spending across 11 providers and 8 coding agents in real time.
TokenTracker counts tokens and spend across Claude Code, Codex, Cursor, Gemini CLI and other agents from local logs and shows them on a local dashboard
A fast Rust tracker of tokens and costs across Claude Code, Codex, Gemini CLI and others that keeps history beyond 30 days.
Token and cost reports for Claude Code, Codex, Gemini CLI and other agents from local logs, by day, month and session
A desktop widget with quota balances and token use across Claude, Codex, Kimi, GLM and other agents in one window.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Tokdash (this card) | no | 92 | |
| ccusage | no | 19,415 | CLI reports from Claude Code logs; Tokdash adds a web dashboard, a terminal mode, quotas and a menu bar app |
| Claude Code Usage Monitor | no | 8,799 | a terminal monitor for Claude Code; Tokdash covers a dozen agents and several machines |
README in short
The README describes four views of the same data, pipx installation and the `tokdash setup` wizard, the client list, optional quota tracking and caveats on cost accuracy and history retention.