start using Agent Browser with your agents
sign up, open a project, and add Agent Browser to its setup script in settings → server.
agent browser is the browser technology for tasks rather than test suites. installed in each conversation's sandbox, it gives agents a managed headless chrome they drive from the command line, on the same machine as the checkout.
what agents do with it
once the setup script has run, agent-browser is in the sandbox's path. agents open a page, work through a flow, and read what the browser actually rendered rather than what the markup implied. that matters for anything behind a login, anything built by javascript, and anything where the answer is what a real user would see. because it runs in the project, the pages it opens and the output it produces stay inside that project.
the matching skill
agent browser ships with a skill of the same name. it carries the practical parts: run flows from the project directory, and remember that browsers and operating system dependencies were installed by the setup script so there is nothing to set up per run. skills load only for the agents you assign them to.
example tasks
- 01
check the signup flow on staging still works end to end.
the agent drives the flow in a managed browser in the conversation's sandbox, reports where it stopped if it stopped, and captures what the page showed at that point.
bashread - 02
the docs page we link to changed. tell me what it says now.
it opens the page in a real browser rather than fetching the raw html, reads the rendered content, and reports the parts that no longer match our copy.
bashgrepedit
what to know before installing
the project needs a sandbox, and node.js must come earlier in the setup script. browsers are memory hungry, so a project doing much of this work wants the standard size or larger.
how to install Agent Browser
Agent Browser goes in the project's setup script, not once per team and not on your laptop.
- give the project a sandbox
Agent Browser runs in each conversation's sandbox, on your team's own sandbox provider account. an owner or admin connects a provider once on the team's providers page, then picks it and a default size for the project, at creation or later in settings → server. a project with no sandbox has nowhere to run it.
- add it to the setup script
settings → server holds the project's setup script, a bash script that runs once on each new sandbox, in /project, with sudo. add Agent Browser's install commands there and save.
- the next sandbox runs it
the script runs when a conversation's sandbox is created or reused, on the conversation's first turn. settings shows the last run with its exit code and output, and a failed run leaves the sandbox usable rather than blocking the project.
- agents use it through bash
from then on every agent in the project reaches Agent Browser with the bash tool, on the same filesystem as the checkout they are working in. the conversation's sandbox keeps it between turns, sleeping when idle and waking with the next one.
where Agent Browser actually lives
each conversation in a project gets its own sandbox on your team's sandbox provider account, and the project's setup script in settings → server decides what is on it. two projects never share a filesystem, so a version pinned for one client cannot collide with another's.
the sandbox provider bills its machine time to your team's account while it runs, and aldena adds nothing. the sizes are on isolated projects, and what agents may run in the sandbox is governed by the project's tool permissions.
faq
what is agent browser?
a command line tool that drives a managed headless chrome. installing it in each conversation's sandbox puts agent-browser on the path and provisions the browser and its operating system dependencies, so agents open and interact with pages without wiring up a test runner first.
how is it different from installing chrome or playwright?
chrome is the browser alone. playwright is a test framework with its own browsers and runner. agent browser is a cli built for driving a browser as a task rather than as a test, which suits checking a live page or working through a flow once.
does it need anything installed first?
yes, node.js. it installs through npm, so node.js has to come earlier in the project's setup script. it then installs the browser and its dependencies with agent-browser install --with-deps.
is there a matching skill?
yes. agent browser ships with a skill of the same name covering running flows from the project directory and the fact that browsers and system dependencies are installed once by the setup script. skills load only for the agents you assign them to.
spin up your first project.
one project per client, product, or piece of work, staffed with intake, requirements, build and code review agents.
