chrome-cdp

name	chrome-cdp
description	Interact with local Chrome browser session (only on explicit user approval after being asked to inspect, debug, or interact with a page open in Chrome)

Chrome CDP

Lightweight Chrome DevTools Protocol CLI. Connects directly via WebSocket — no Puppeteer, works with 100+ tabs, instant connection.

Prerequisites

• Chrome (or Chromium, Brave, Edge, Vivaldi) with remote debugging enabled: open chrome://inspect/#remote-debugging and toggle the switch
• Node.js 22+ available through nvm on this machine. Use the wrapper scripts/cdp, which enters Node 22 explicitly before running the CLI.
• If your browser's DevToolsActivePort is in a non-standard location, set CDP_PORT_FILE to its full path

Commands

All commands use scripts/cdp. The <target> is a unique targetId prefix from list; copy the full prefix shown in the list output (for example 6BE827FA). The CLI rejects ambiguous prefixes.

List open pages

scripts/cdp list

Take a screenshot

scripts/cdp shot <target> [file]    # default: screenshot-<target>.png in runtime dir

Captures the viewport only. Scroll first with eval if you need content below the fold. Output includes the page's DPR and coordinate conversion hint (see Coordinates below).

Accessibility tree snapshot

scripts/cdp snap <target>

Evaluate JavaScript

scripts/cdp eval <target> <expr>

Watch out: avoid index-based selection (querySelectorAll(...)[i]) across multiple eval calls when the DOM can change between them (e.g. after clicking Ignore, card indices shift). Collect all data in one eval or use stable selectors.

Other commands

scripts/cdp html    <target> [selector]   # full page or element HTML
scripts/cdp nav     <target> <url>         # navigate and wait for load
scripts/cdp net     <target>               # resource timing entries
scripts/cdp click   <target> <selector>    # click element by CSS selector
scripts/cdp clickxy <target> <x> <y>       # click at CSS pixel coords
scripts/cdp type    <target> <text>         # Input.insertText at current focus; works in cross-origin iframes unlike eval
scripts/cdp loadall <target> <selector> [ms]  # click "load more" until gone (default 1500ms between clicks)
scripts/cdp evalraw <target> <method> [json]  # raw CDP command passthrough
scripts/cdp open    [url]                  # open new tab (each triggers Allow prompt)
scripts/cdp stop    [target]               # stop daemon(s)

Coordinates

shot saves an image at native resolution: image pixels = CSS pixels × DPR. CDP Input events (clickxy etc.) take CSS pixels.

CSS px = screenshot image px / DPR

shot prints the DPR for the current page. Typical Retina (DPR=2): divide screenshot coords by 2.

Tips

• Prefer snap --compact over html for page structure.
• Use type (not eval) to enter text in cross-origin iframes — click/clickxy to focus first, then type.
• Chrome shows an "Allow debugging" modal once per tab on first access. A background daemon keeps the session alive so subsequent commands need no further approval. Daemons auto-exit after 20 minutes of inactivity.