Docs

Page commands

Inspect and interact with the current page DOM from the shell.

Page commands run scripts in the active tab to read and manipulate the DOM. They only execute when you type them.

Discovery

links           # all links with href
hints           # Vimium-style link hints (or press f on page)
inputs          # form fields
images          # images with src/alt
meta            # meta tags
read            # page text content
seek query      # find on page from shell
tech            # detect frameworks
reqs            # network requests
frames          # iframes
viewport        # viewport dimensions
jsonld          # structured data

For keyboard-driven hints and scroll without the terminal, see Page hotkeys.

Interaction

link 3          # follow link #3 (after links)
input focus 2   # focus field #2 (after inputs)
fill 1 "text"   # fill field #1
press Enter     # press a key
click 1         # click element #1
pick            # interactive element picker
scroll down     # scroll page
zoom 1.2        # zoom level
volume mute     # page audio
shot            # screenshot to clipboard

Dev inspection

tech            # React, Vue, Next.js, etc.
reqs            # XHR/fetch requests
reqs | grep api # filter requests
perf            # performance timing
env             # page environment
storage         # localStorage/sessionStorage
cookies         # page cookies
frames          # iframe listing

Typical workflows

links > /notes/page-links.txt

Fill a login form

inputs
fill 1 "user@example.com"
fill 2 "password"
press Enter

Debug API calls

reqs | grep -E "4[0-9]{2}|5[0-9]{2}"

Screenshot for bug report

shot
notify "Screenshot copied"

Context pattern

List commands set context for numbered follow-ups:

links           # sets link context
link 1          # uses context
links           # refreshes context
click 3         # different context
Warning
Page commands inject scripts via chrome.scripting. They run only on your command. Some sites with strict CSP may block injection.