Timeslicer

Timeslicer reads text from your active window, groups the fragments that belong to the same piece of work, and labels the group. The result is a task name such as "billing webhook retries bug" instead of an app name like "VS Code." It works without rules or timers, which is the main difference from tools that sort activity into fixed categories.

The app sorts a page into focus or distraction using Google Gemini and can block distracting content in real time. You define your own blocklists and schedules, and classification runs against what you are looking at rather than a fixed list of domains. This is the part the company markets as a contextual distraction blocker.

Beyond app and window titles, Timeslicer can read up to 500 characters of page text, scan AI coding assistant transcripts, and capture text inside desktop apps like Slack, Discord, Notion and Obsidian. Page text is scrubbed for emails, card numbers and ID numbers before it leaves the tab, and keys and tokens are stripped from coding prompts first.

Your task history lives on your machine in plain SQLite and JSONL files in the app data folder. You can open them, export them or delete them without asking. The site publishes four terminal commands to verify local storage, count every outbound labeling call, and confirm that a pause really stops recording.

Settings has an Integrations tab that connects Jira, Linear, GitHub or Attio. Timeslicer reads the tickets assigned to you and matches a sitting of work to an open ticket, showing it in the block's detail view with a link. It writes nothing back to your tracker, and disconnecting deletes the token and copied tickets locally.

On the team plan, a daily summary is locked on your computer with your team's key before it is sent, and Timeslicer's servers pass the locked bundle along without holding a key. Managers see task names, active time and block timing, but never screenshots or raw screen text, and they cannot see anything older than the last 7 days.