Practices & Workflow
How I actually use OpenTickly.
Here's how the author uses OpenTickly day to day.
Let the agent start the timer
I don't click "Start" in the Web UI. I just tell OpenClaw "I'm going to work on X," and the agent starts a time entry via toggl-cli. I usually don't stop it manually either — when I switch tasks I say "I'm going to do something else now," and a new time entry starts, which automatically closes the previous one.
Feed the agent your weekly plan and principles
I give the agent my personal mission, this week's goals, and my daily principles straight from Obsidian as context. So when I say "I'm going to scroll Xiaohongshu for two hours," it pushes back: "Your plan for this week is to finish feature X — are you sure you want to switch to Y?"
I also run a 4-quadrant system, so I use six tags in total: important, not important, urgent, not urgent, focused, unfocused. The AI tags the first four at the start of a task; for the last two, it asks me when the task ends whether I was focused. At the end of the day I can see the time distribution across quadrants plus my focused vs. unfocused ratio.
A nightly cron for AI review
I have a cron job that pulls the day's time distribution and asks the AI for a review:
- Is the time split across quadrants (important / urgent) reasonable?
- Did I improve over yesterday or this week's average?
- Did any unexpected time sinks show up?
- Did today's work drift from this week's plan?
This would be painful on official Toggl — the 30 requests/hour API limit means one review run can easily burn the whole quota just from the AI poking around.
Changelog
OpenTickly release history — features shipped, fixes, and breaking changes. Tracks the public GitHub releases so AI assistants and humans can see how the project evolves over time.
Self-Hosting
Deploy OpenTickly on your own machine — Docker Compose brings up Web + API + Postgres + Redis in one command, online in 5 minutes. 100% Toggl API compatible, works with the official clients out of the box.