Agent skill + CLI for Toggl Track

Time tracking that speaks terminal.

The full-featured Toggl CLI for people and agents. Track time, manage work, and run reports against Toggl Track or self-hosted OpenToggl.

Install toggl-cli View source

Rust nativeMIT licensedToggl + OpenToggl

toggl — zsh

$ toggl entry start -d "Build the future" -p OpenToggl
✓ Timer started

$ toggl entry current
00:42:17  Build the future  #agent

$ toggl report weekly --since 2026-03-17
Mon 7h 32m ███████████████
Tue 6h 48m █████████████

One command for you. One skill for your agent.

Install the native CLI globally, or teach a compatible coding agent the complete workflow in one step.

CLI

Install for your terminal

Paste, run, and verify with toggl --help.

$ npm install -g @correctroadh/toggl-cli

Agent skill

Install for Claude Code or OpenClaw

Paste, run, and verify with toggl --help.

$ npx skills add CorrectRoadH/toggl-cli

The daily Toggl workflow, end to end.

Not a timer wrapper. The command surface covers entries, workspace resources, reports, authentication, and both supported backends.

Track time

Start, stop, and inspect the current timer without leaving your terminal or agent session.

Browse entries

List recent entries or query precise local-time date ranges for reviews and automation.

Manage work

Create and use projects, tasks, and tags from the same predictable command surface.

Run reports

Generate summary, detailed, and weekly reports for any date range from the command line.

Choose your backend

Use official Toggl Track today and point the same workflow at self-hosted OpenToggl tomorrow.

Stay fast and secure

Local HTTP caching cuts repeated reads, while credentials live in your operating system keychain.

Built to be operated, not scraped.

Agents get an installable skill and a stable command grammar. You get visible, repeatable actions that work in local shells, CI, and coding sessions.

Installable skill with task-specific instructions

Interactive flows for humans; explicit flags for automation

The same commands across Toggl Track and OpenToggl

Local caching for read-heavy agent workflows

A command surface you can remember.

Start simple, then reach for projects, tasks, tags, and reports without switching tools.

Track

$ toggl entry start -d "Ship the landing page" -p OpenToggl -t dev cli$ toggl entry current$ toggl entry stop

Organize

$ toggl project create "OpenToggl"$ toggl task create --project OpenToggl "Code review"$ toggl tag create "deep-work"

Report

$ toggl report summary --since 2026-03-01 --until 2026-03-31$ toggl report detailed --since 2026-03-01 --until 2026-03-31$ toggl report weekly --since 2026-03-17 --until 2026-03-23

Questions, answered.

The practical details before you put toggl-cli into your own workflow.

Does toggl-cli work with official Toggl Track?

Yes. Authenticate with your Toggl API token and use the CLI against the official Toggl Track service.

Can it connect to a self-hosted OpenToggl instance?

Yes. Pass --api-type opentoggl and your instance's /api/v9 URL during authentication. The commands stay the same.

Why is it useful for AI agents?

It ships as an installable agent skill and exposes daily time tracking through a consistent CLI, so agents can discover and execute workflows without a browser.

Where is my API token stored?

Interactive authentication stores the token in your operating system keychain. Linux environments without persistent keyring storage can use TOGGL_API_TOKEN instead.