Browser Use is the open-source framework for AI agents that control a browser. Every Browser Use Cloud session now ships with its own AgentMail inbox, so agents can sign up, verify, and stay logged in with no human in the loop.
Company: Browser Use, the open-source framework that lets AI agents control a web browser
Problem: Browser agents get stuck the moment a site asks them to check an email, signup confirmations, OTP codes, magic links, 2FA, password resets. Without an inbox of its own, the agent stalls and a human has to step in.
Result: Every Browser Use Cloud session is pre-loaded with its own AgentMail inbox. Zero config, no API key. Agents sign up for services, pull verification codes, and keep a persistent identity across sessions, fully autonomously.
Stack: AgentMail + Browser Use
What Browser Use Is
Browser Use is the most widely used open-source framework for giving AI agents control of a web browser. It makes websites readable and actionable for agents so they can navigate, click, fill forms, and complete real tasks instead of stopping at the edge of an API.
It started as a weekend experiment by Magnus Müller and Gregor Žunič out of ETH Zurich's Student Project House, shipped to Hacker News in four days, and grew into one of the fastest-rising projects in the agent ecosystem: 96,000+ GitHub stars, with many Fortune 500 and YC teams among its users. The company raised a $17M seed led by Felicis, with participation from Paul Graham, A Capital, Nexus Venture Partners, and Y Combinator. Browser Use Cloud is their hosted product, where developers run agent sessions without standing up their own browser infrastructure.
The Problem
Most of the web is locked behind an email address. Sign up for a service and you need an email. Confirm the account and you have to open your inbox. Reset a password, clear a 2FA prompt, pick up a receipt, all of it routes back through email.
That is the one wall a browser agent can't climb on its own. Browser Use could already drive the browser better than anything on the market, but the moment a site said "check your email to continue," the run stopped. The agent had no inbox to receive the confirmation link or the OTP, so the task died on a verification screen and a human had to jump in to paste a code.
The constraint wasn't browser control. It was identity. To finish the long tail of real web tasks like account creation, authenticated flows, anything behind a login, every session needed an email address it actually owned. And it needed one without the developer wiring up an email provider, a webhook listener, and an inbound parser for every agent.
Why AgentMail
AgentMail gives agents real inboxes as a primitive: call the API, get an address, send and receive from day one. That maps directly onto what a browser agent is missing at the verification screen, so instead of a developer-side integration, the two teams built it straight into the session.
Every Browser Use Cloud session now comes with its own AgentMail inbox automatically. No API key, no configuration, nothing to set up. Give the agent a task like "sign up for Instagram" and it creates an inbox, fills in the address, submits the form, waits for the verification email, confirms the account, and keeps going. Authentication codes sent mid-session land in the agent's inbox and get picked up on their own. No pauses, no tab-switching, no human in the loop.
The inbox also gives the agent a persistent identity. With persistent profiles, an agent that gets logged out can log back in using the same email, across sessions, without a person re-entering anything. Each session gets its own isolated inbox; when the session ends, the inbox goes with it, so every run can start clean.
"Your agent is naked. Every browser agent needs an email inbox. When you sign up, you need an email. When you verify, you need an email. When you send a message, you need an email. Now every Browser Use session comes pre-loaded with an inbox, so the agent can sign up, authenticate, and keep going without a human in the loop. This isn't our API. It's the future."
"Email is a natural next step for agents that already get real work done in the browser, and AgentMail made it effortless to add. The team is great to work with, genuinely fun, and incredibly responsive. Every session now comes with its own inbox, so our agents can handle signup and verification on their own."
Magnus Müller, Co-founder and CEO, Browser Use
What This Unlocks
With an inbox in every session, browser agents can finally complete the flows that used to require a person:
Account creation. Sign up for any service that needs email verification. The agent manages the address, the confirmation link, and everything after it.
2FA and OTP flows. Authentication codes sent during a session route straight to the agent's inbox and are picked up automatically, with no interruption.
Confirmation and receipt emails. The agent reads what it receives, not just what it sends. Booking confirmations, order receipts, and access links get processed in real time.
Research and competitive monitoring. Trial signups, newsletter subscriptions, and product registrations all run to completion. The agent signs up, gets access, and reports back.
How Developers Use It
For Browser Use Cloud, there is nothing to do. Open cloud.browser-use.com and the inbox is already there.
For the open-source framework, developers connect AgentMail directly with an API key. The integration lives in Browser Use's official examples under examples/integrations/agentmail/ and adds two tools to the agent: one to get the inbox address, one to fetch the latest email.
import asyncio
import os
os.environ["BROWSER_USE_API_KEY"] = "your-browser-use-key"
os.environ["AGENTMAIL_API_KEY"] = "your-agentmail-key"
from browser_use import Agent, ChatBrowserUse, Browser
from agentmail import AsyncAgentMail
from agentmail.inboxes.types import CreateInboxRequest
from email_tools import EmailTools # Provided in the Browser Use example
async def main():
client = AsyncAgentMail(api_key=os.environ["AGENTMAIL_API_KEY"])
inbox = await client.inboxes.create(request=CreateInboxRequest())
email_tools = EmailTools(client, inbox=inbox)
browser = Browser(use_cloud=True)
agent = Agent(
task=f"Sign up for Reddit using {inbox.email} and verify the account.",
llm=ChatBrowserUse(model="bu-2-0"),
browser=browser,
tools=email_tools,
)
await agent.run()
if __name__ == "__main__":
asyncio.run(main())
In Production
Every Browser Use Cloud session now ships with an AgentMail inbox, across a user base built on 96,000+ GitHub stars, with many Fortune 500 and YC teams among its users. The verification wall that used to stop a browser agent is gone by default.
Products used: Programmatic inboxes, Send and Receive, Webhooks, Threaded conversations, structured extraction of OTPs and verification links, persistent identity across sessions.
Building something similar? console.agentmail.to/sign-up
AgentMail gives your agents real inboxes. Create inboxes via API. Send and receive Emails with 0 complexity. Free to start.

