Media
1 item
Description
Work with your company's Lleverage agent beside the page you are on: it can read the page and fill in forms with you. Your Lleverage agent, beside the page you are working on. A lot of back-office work still happens in web pages: a supplier portal, a customer's order screen, the web client of the ERP, a form that wants the same twenty fields typed in again. The Lleverage extension opens your company's Lleverage agent in a side panel next to that page, so the agent can work in it with you. What it does • Reads the page the way a screen reader does: headings, tables, rows, fields and their current values. Every message also carries a screenshot, so "this one" means what you are looking at. • Fills in fields, picks options, ticks boxes, clicks, presses keys and scrolls, in your own signed-in browser. After every step it reads the page back to check what happened. • Works in a call too. A voice call shares the attached tab straight away, so you can talk the agent through the page. You stay in charge • The agent works only on the page you attached with the Lleverage button. Switching tabs doesn't move it. • Anything that can't simply be taken back waits for you: submitting a form, sending, paying, deleting, publishing, approving, ordering, or going to another site. An Allow card appears in the chat and you decide. Searches, filters and saving go through without asking. • Other sites are allowed one at a time, only when you choose "Allow on …". • It never reads or fills passwords, card numbers or one-time codes. Getting started 1. Install the extension and pin it to the toolbar. 2. Open the page you want to work on and press the Lleverage button, or Alt+Shift+L. 3. Sign in with your Lleverage account, then ask the agent what is on the page or what to fill in. The panel is the Lleverage app's own agent chat, so your sign-in, conversations and projects are the ones you already have. A Lleverage account is required. What it doesn't do yet Content inside frames from other sites, closed shadow roots and canvas-based interfaces are out of reach. A few pages accept only keys typed by a person and ignore the agent's key presses.