Skip to main content

Browser

The agent drives a cloud browser in a live Computer side panel. Watch it work. Take over to log in. Then let it extract information or capture screenshots.

Enable Browser on Ordify Productivity in Edit Agent → Integrations. You do not connect a third-party account or complete OAuth.

What you can do​

  • Open JavaScript-heavy pages and work through multi-step sites
  • Take over to complete login, SSO, or 2FA on sites such as LinkedIn, Gmail, or vendor portals — your credentials stay in the browser
  • Extract information after you hand the session back
  • Screenshot pages; shots appear in chat and in the Files strip

How it works​

  1. Enable Browser on Ordify Productivity in Integrations.
  2. Ask the agent in chat to open a page or work through a site.
  3. The Computer panel opens with a live view you can watch.
  4. If a login is needed, take over, finish it yourself, then click Hand back.

The agent can click, type, navigate, extract information, and screenshot pages. Collapse the panel to stay in chat. Reopen it from the Browser control in the tool trace.

Getting started​

  1. Find the agent in your Library or sidebar and click Edit.
  2. Open the Integrations tab.
  3. Under Apps, tick Ordify Productivity. It is marked Built-in — there is no connect step.
  4. On the right, tick Browser. Leave tools this agent does not need off.
  5. Click Save Changes.
  6. Open chat and ask the agent to open a page.

Watch and take over​

Take over lets you drive the browser yourself — the usual way to complete login, SSO, or 2FA. Your credentials stay in the browser. They are never sent to the agent or typed into chat. Never paste passwords into chat.

Status in the panel:

  • Needs you — the agent is waiting for you to log in or finish a step only you can do.
  • You control — you have taken over and are driving the browser.
  • Hand back — you are done. Click it so the agent continues in the same session.

After you hand back, the agent can extract information or capture screenshots in that signed-in session.

Stay signed in​

After a successful login, later visits to the same site often stay signed in. If the site asks again (common on LinkedIn), take over and sign in again.

Screenshots and files​

The agent can screenshot the page it is viewing. Screenshots land in the conversation and in the Files strip in the Computer panel.

Use this when a table, pricing page, or portal view is easier to share as an image than as extracted text.

Search vs Browser​

Search finds current titles, links, and snippets. Browser opens a real page.

  • If the agent only needs to look something up, give it Search.
  • If it needs a spreadsheet or Word file from what it finds, use Excel and spreadsheets or Word and documents.
  • If it needs to click through a site, wait while you log in, extract from a live page, or take a screenshot, use Browser.

Search does not log in, click, or take screenshots.

Limits​

The browser is for reading, extracting, and screenshots. Do not expect it to submit forms or change data on sites.

Browser is for chat. Unattended jobs cannot complete a login handoff.

Example​

Ask in chat after Browser is enabled:

  • "Open this pricing page and screenshot the comparison table."
  • "Go to LinkedIn, wait while I log in, then pull the latest posts from this company page."
  • "After I log in, extract the open invoices from this vendor portal."

Use cases​

  • Research: Open a competitor pricing page, screenshot the comparison table, and pull the figures into a follow-up.
  • Sales: Take over to sign in to LinkedIn, then let the agent extract the latest posts from a company page.
  • Operations: After you log in to a vendor portal, have the agent extract open invoices from the live page.

What's next​