Persistent browser infrastructure

Persistent browsers that branch with your work.

Open a browser once. Keep its logins, tabs, task sessions, and workflows. Fork it into an independent browser when the work needs to split.

Use the CLI Install the agent skill See the browser

Browser command line

Create once. Fork when work branches.

webbrain new · fork · run

terminal

# Start a persistent browser and make it current
webbrain new research

# Work in its visible browser
webbrain run "Research competitor pricing"

# Pause for a consistent profile, tab, and task-session copy
webbrain pause
webbrain fork research --name eu-research --pause

# Continue the copied context independently
webbrain continue eu-research "Focus on Europe"

import os
from clients.python.webbrain_client import WebBrainClient

client = WebBrainClient(os.environ["WEBBRAIN_API_KEY"])
session = client.create_browser_session()
ready = client.wait_for_browser_session(session["id"])

run = client.create_run(
    ready["id"],
    "Open github.com and summarize the trending repos",
)

result = client.wait_for_run(ready["id"], run["run_id"])
print(result["result"])

export WEBBRAIN_API_KEY='wbp_your_key_here'

# 1. Create a browser
SESSION_ID=$(curl -sS -X POST https://webbrain.cloud/api/browser-sessions   -H "Authorization: Bearer $WEBBRAIN_API_KEY"   -H "Content-Type: application/json"   -d '{"type":"normal"}' | jq -r '.browser_session.id')

# 2. Wait for the extension bridge
until [ "$(curl -sS   "https://webbrain.cloud/api/browser-sessions/$SESSION_ID"   -H "Authorization: Bearer $WEBBRAIN_API_KEY"   | jq -r '.browser_session.runtime_ready')" = "true" ]; do sleep 2; done

# 3. Start a visible run
RUN_ID=$(curl -sS -X POST   "https://webbrain.cloud/api/browser-sessions/$SESSION_ID/runs"   -H "Authorization: Bearer $WEBBRAIN_API_KEY"   -H "Content-Type: application/json"   -d '{"task":"Open github.com and summarize the trending repos"}'   | jq -r '.run_id')

# 4. Read the result
curl -sS   "https://webbrain.cloud/api/browser-sessions/$SESSION_ID/runs/$RUN_ID"   -H "Authorization: Bearer $WEBBRAIN_API_KEY" | jq

<?php

require_once __DIR__ . '/clients/php/WebBrainClient.php';

$client = new WebBrainClient(getenv('WEBBRAIN_API_KEY') ?: '');
$session = $client->createBrowserSession();
$ready = $client->waitForBrowserSession($session['id']);

$run = $client->createRun(
    $ready['id'],
    'Open github.com and summarize the trending repos',
);

$finished = $client->waitForRun($ready['id'], $run['run_id']);
print_r($finished['result']);

The WYSIWYG programmable browser in the cloud.

What you see is what you get. A full browser you and your agent can both use. Watch the page, handle a manual step, and run your next task through the API — all in the same session.

WebBrain running beside a GitHub issue in Chrome, with the browser tabs, address bar, and agent conversation visible. — WebBrain interface shown in Chrome. Cloud sessions make the browser available remotely through your dashboard. View full size.

See the whole browser.

Open the live browser from your dashboard. See the page, tabs, and WebBrain interface together, with the context you need to understand a run.

Take the controls.

Sign in, navigate, or handle a step yourself in the remote browser. Your agent uses that same session when you start its next task.

Branch the work.

Fork a paused browser into an independent copy of its profile, tabs, and completed task sessions. Turn repeated tasks into saved workflows.

A few things worth knowing.

How the browser works, what stays with your session, and what you pay for.

What does WYSIWYG mean here?

WYSIWYG means “what you see is what you get.” In WebBrain Cloud, you can see and interact with the full graphical browser your agent uses, including its tabs and WebBrain interface. Manual actions and API tasks happen in the same browser session.

Why can WebBrain Cloud cost more than other browser automation services?

WYSIWYG is our main differentiator. We run a full graphical browser, live desktop streaming, and the WebBrain agent together. That experience needs more CPU and memory than a lightweight headless setup, so we use a higher-spec server configuration. The price reflects the infrastructure behind a browser you can both see and control. See our pricing page for current rates and usage details.

Can I use it entirely from code?

Yes. Create a browser session, send tasks, read results, and replay saved workflows through the API. Use Node.js, Python, PHP, or REST. The visual browser is there whenever you want to inspect the session or handle a manual step. Start with the quickstart.

Will my browser remember its logins?

Normal browsers keep their profile on persistent storage. Pause and resume to return to the same browser, or fork a paused browser to create an independent copy with the same logins, tabs, and completed task context. Incognito browsers are discarded when destroyed.

How is Cloud different from the WebBrain extension?

The WebBrain extension runs in your own browser. WebBrain Cloud hosts the browser remotely and adds an API, session lifecycle controls, and saved workflows. You get the WebBrain interface in a browser that your code and agents can access in the cloud.

Give your next workflow a browser you can see.

Build your first workflow View pricing