| name | clawchrome-cli |
| description | Use when controlling or inspecting a browser with clawchrome-cli from a shell, including navigation, accessibility snapshots, clicking and filling elements by snapshot refs, screenshots, tabs, waits, form controls, keyboard input, and common browser workflows. |
clawchrome-cli
Use clawchrome-cli to drive a browser from the shell. The normal loop is: open or select a page, take a snapshot, act on @eNN refs from that snapshot, then snapshot again.
Quick Start
clawchrome-cli open https://example.com
clawchrome-cli snapshot
clawchrome-cli click @e12
clawchrome-cli fill @e17 "search terms"
clawchrome-cli press Enter
clawchrome-cli screenshot ./page.png
Element commands use snapshot refs such as @e12 or @e27. Run clawchrome-cli snapshot after navigation or page changes to get current refs.
Help
- Run
clawchrome-cli --help for the full command list.
- Run
clawchrome-cli <command> --help or clawchrome-cli <command> -h before guessing flags.
- Subcommand help shows the required args, available flags, and examples for that specific command.
Main Commands
open <url>: navigate the current page and print a snapshot.snapshot [--form] [--text] [--full]: inspect the current page. Use --form for controls, --text for readable text, and --full when truncation hides needed refs.click @eNN: click a link, button, checkbox, or other interactive element.fill @eNN <text>: fill an input or textarea.type <text>: type into the currently focused element.press <key>: press keys such as Enter, Tab, Escape, or ArrowDown.scroll <up|down|top|bottom>: move around the page.wait <ms|text>: wait for a duration or for text to appear.screenshot [path]: save a screenshot. Without a path, the tool chooses a temp file and prints it.pages: list open tabs/pages.newpage <url>: open a new tab.selectpage <id>: switch tabs by ID from pages.closepage <id>: close a tab; the last page cannot be closed.back, forward, reload: browser history and reload actions.hover @eNN and drag @eNN @eNN: pointer interactions by snapshot ref.fillform @eNN=<value>...: fill several fields in one command.form clear|check|uncheck|select|upload ...: targeted input clearing, checkbox, select, and file upload helpers.dialog accept|dismiss [text]: handle alerts, confirms, and prompts.resize <width> <height>: set viewport size.video start [path] and video stop: record and stop page video.start [url] [--token <token>] [--agent-name <name>]: start the browsing session. --token saves HTTP auth for later commands; --agent-name saves agent metadata for runtime HTTP requests.stop: stop the bridge.version: print the installed version.self-update: update the CLI.
Common Workflows
Navigate and act from the snapshot:
clawchrome-cli open https://example.com/login
clawchrome-cli snapshot --form
clawchrome-cli fill @e12 "user@example.com"
clawchrome-cli fill @e13 "correct horse battery staple"
clawchrome-cli click @e14
clawchrome-cli wait "Dashboard"
Work with tabs:
clawchrome-cli pages
clawchrome-cli newpage https://example.com/docs
clawchrome-cli selectpage 2
clawchrome-cli closepage 1
Capture page state:
clawchrome-cli snapshot --full
clawchrome-cli screenshot
clawchrome-cli screenshot ./checkout.png --full-page
clawchrome-cli screenshot ./checkout.jpeg --full-page --format jpeg
Configure HTTP auth:
clawchrome-cli start --token "$CLAWCHROME_TOKEN" --agent-name codex-worker
clawchrome-cli status
Handle forms and dialogs:
clawchrome-cli form clear @e17
clawchrome-cli form check @e18
clawchrome-cli form select @e19 "United States"
clawchrome-cli form upload @e20 ./avatar.png
clawchrome-cli dialog accept
