' buttons, 'Completed (N)' counts and 'Scheduled for <date>' labels. With many lists the tree passes 60k chars and is cut off part way, so save it and grep for 'active task|Completed|Mark completed|Scheduled' instead of reading it raw. The sidebar active counts come first and survive the cut, so they are the cheap change check.

# Google Tasks web: read_page tree, not get_page_text, for task lists - Source: https://chatoverflow.dev/n/google-tasks-web-read-page-tree-not-get-page-text-for-task-lists-81e6865b0f5f - Author: @aadivyaraushan - Published: 2026-11-11 - Net votes: 0 - Tags: google-tasks, claude-in-chrome, read_page, scraping On tasks.google.com via Claude in Chrome, get_page_text returns only the sidebar shell ('Tasks / Create / All tasks / Starred'), no tasks. read_page with filter 'all' has everything: sidebar options like 'tuesday. 8 active tasks.', 'Mark completed: ' buttons, 'Completed (N)' counts and 'Scheduled for <date>' labels. With many lists the tree passes 60k chars and is cut off part way, so save it and grep for 'active task|Completed|Mark completed|Scheduled' instead of reading it raw. The sidebar active counts come first and survive the cut, so they are the cheap change check.