Browser Automation

Write JavaScript here to open, read, click, type in and screenshot your other tabs, through a browser extension you install and approve.

Desktop Chrome or Edge only. The extension is loaded unpacked (Developer mode) and asks you to approve each tab that connects.

Checking for the extension…

Download extension (v0.1.0, .zip)

How to install it

Download the ZIP above and unzip it somewhere you'll keep it (Chrome loads it from that folder each time). You'll get a rafb-automation folder.
Open chrome://extensions (or edge://extensions) in a new tab chrome://extensions. Web pages can't link there, so paste it into the address bar.
Turn on Developer mode (top right), click Load unpacked and choose the rafb-automation folder.
Optional: click Details on the extension and turn on Allow User Scripts. Then eval() runs without Chrome's "started debugging this browser" bar.
Come back here and click Connect. The extension opens a small window asking you to allow this tab.

To update: download the new ZIP, replace the folder's contents, and click the reload arrow on the extension's card. The extension's source is in the automation/extension folder of the rafb.me repository; everything in the ZIP is readable JavaScript.

Script Sign in to save named scripts. Your draft is kept in this browser.

Name

Output

Script API reference

Tabs

await tabs.open(url, { active }) → tab (waits for it to load)

await tabs.list() → every tab

await tabs.find(match) → first tab whose URL or title matches a string or RegExp

await tabs.get(id)

tab.id, tab.url, tab.title (as of the last call)

await tab.goto(url) / reload() / back() / activate() / close()

Reading

await tab.text(sel) / texts(sel) / html(sel) / value(sel)

await tab.attr(sel, name) / attrs(sel, name)

await tab.count(sel) / exists(sel)

await tab.links(sel?) → [{ text, href }]

await tab.table(sel) → rows (objects keyed by the header row)

await tab.eval(fn, ...args) → runs fn in the page, returns its (JSON) result

await tab.waitFor(sel, { state, timeout }) state: visible | attached | detached | enabled

Acting

await tab.click(sel, { trusted })

await tab.fill(sel, value) (inputs, textareas, selects, checkboxes)

await tab.type(sel, text, { trusted }) appends text

await tab.press('Enter' | 'Ctrl+A' | 'Tab Tab', sel?) (always trusted)

await tab.screenshot({ fullPage }) → Blob (shown below)

await tab.pdf({ landscape }) → Blob

Output and helpers

log(...values), table(rows), show(blob), download(name, data)

await sleep(ms), stopped(); return a value to print it

Selectors

CSS (also inside shadow DOM), text=Sign in, label=Email, xpath=//h1

Actions wait up to 10 seconds ({ timeout } changes it) for the element to be visible and enabled. trusted: true clicks and types through Chrome's debugger, so the page sees real user input; that, screenshots, PDFs and (without user scripts) eval show Chrome's debugging bar while they run.

🤖Browser Automation

Browser Automation