start using Google Chrome with your agents
sign up, open a project, and add Google Chrome to its setup script in settings → server.
some pages are html and some pages are a program that produces html. installing chrome in each conversation's sandbox gives agents a real rendering engine for the second kind, on the same machine as the checkout.
what agents do with it
once the setup script has run, google-chrome-stable is in the sandbox's path. agents drive it headless through the bash tool: render a url, capture a screenshot, print a page to pdf, or run the app's own headless rendering step. the output lands on the sandbox's filesystem, so the agent reads it back and acts on it without anything leaving the project.
when chrome is the right choice
chrome on its own suits projects that already drive a browser their own way: a language binding, a build step that snapshots a page, a script that generates a pdf. if the project has an end to end suite, the playwright technology is the better fit because it brings its own browsers and runner. if agents need a general purpose browsing tool with a cli, agent browser is built for that.
example tasks
- 01
the marketing page renders blank for some users. check what chrome actually gets.
the agent renders the page headless in the conversation's sandbox, captures the result, and reads the console output to find the script that failed before the content mounted.
bashread - 02
generate the pdf export and confirm the layout survived the last css change.
it prints the page to pdf in the conversation's sandbox, checks the output, and iterates on the print stylesheet until the layout holds.
basheditgrep
what to know before installing
the project needs a sandbox, and a browser is memory hungry, so a project doing much rendering wants the standard size or larger. chrome has no matching skill of its own, so the guidance for browser work lives with the playwright and agent browser skills instead.
how to install Google Chrome
Google Chrome goes in the project's setup script, not once per team and not on your laptop.
- give the project a sandbox
Google Chrome 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 Google Chrome'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 Google Chrome 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 Google Chrome 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 does installing chrome on a project give the agents?
google chrome stable in each conversation's sandbox, driven headless through the bash tool. agents render a page, capture a screenshot or a pdf, and check what the browser actually produced rather than what the html suggests it should.
how is this different from the web fetch tool?
web fetch retrieves a url and summarises it. chrome renders it, which is the difference that matters for anything built by javascript. a page that is empty html until a script runs is invisible to a fetch and visible to a browser.
is chrome the same as installing playwright?
no. playwright brings its own browser binaries and a test runner. chrome is the browser on its own, for cases where the project already drives it, for example through a language binding or a headless rendering step in the build.
does it run headed?
each conversation's sandbox has no display, so chrome is used headless. screenshots and pdfs are written to the sandbox's filesystem, where agents read them back with the same tools they use for any other file.
spin up your first project.
one project per client, product, or piece of work, staffed with intake, requirements, build and code review agents.
