Introducing the agent web browser: faster, cheaper & more reliable
Tasklet already works great with thousands of sites via direct integrations. But when a website doesn’t offer an API or integration, things get awkward. We offered Computer Use — spin up a full cloud VM and let the agent drive it like a human — but a VM is slow to start, expensive to run, and overkill for most web tasks.
Today we’re announcing that every Tasklet agent now has a built-in browser. It’s fast, more cost efficient, and always ready — no setup required.
Fast, powerful web automation
The built-in browser is a full Chrome instance your agent controls directly. There’s no VM to boot and no connection to configure — it’s just there. It can do anything you can do: navigate pages, click buttons, fill out forms, extract data, upload and download files, and take screenshots. It uses far fewer credits than Computer Use, and the agent reaches for it naturally when a direct integration isn’t available.
You stay in control
Tasklet will never ask for your password. When a site requires a login, the agent surfaces a live browser preview right in the conversation. You take over, handle authentication however you normally would, and hand control back. The agent resumes exactly where it left off. You and the agent share the browser together, so you can jump in at any point to help — or just watch it work.
Log in once, stay logged in
The browser keeps a persistent profile for each agent. Sign into a site once, and the agent stays signed in for every future task — no re-authentication required. Sessions, cookies, and site preferences all carry over.
Try it
- Log into my Google Ads account and take a screenshot of this week’s campaign performance dashboard
- Download receipts from my Gmail and upload them to my expense tracking system
- Every Monday, reconcile my QuickBooks entries against my bank transactions and flag any mismatches
- Log into our Shopify admin, find all orders from this week that haven’t shipped yet, and send me the list
- Go to our WordPress admin and publish the three blog posts that are sitting in draft
- Log into my company’s HR portal, download my pay stubs and tax documents, and file them in Dropbox
- Go to DoorDash and re-order my last Thai food order
The whole web is now in reach
APIs cover a lot of ground, but plenty of the most important tools in any workflow — vendor portals, legacy systems, government sites, dashboards with no export button — don’t have one. With the built-in browser, that’s no longer a blocker. If a human can access it through a browser, your agent can too.
Have feedback? Drop me a note at michael@tasklet.ai