Browser

Drive a real Chromium with the agent-browser CLI. Use it for anything a page renders: local dev apps, JS-heavy sites, logins, scraping, screenshots, console/network debugging.

The Core Loop (this is law)

start-browser <url>              # first use only: starts Xvfb + Chromium + dashboard
agent-browser snapshot -i        # interactive elements with @eN refs
agent-browser click @e3          # act via refs
agent-browser wait --url "**/dashboard"   # semantic wait, never bare sleeps
agent-browser snapshot -i        # ALWAYS re-snapshot after the page changes

Refs go stale the moment the page changes (navigation, submit, dialog, re-render). Acting on a stale ref is the #1 failure — re-snapshot first. Always use -i (interactive-only) to keep output small; add -u to include link URLs.

Environment facts:

• Nothing is running at startup — call start-browser [url] once, then reuse the same session for everything.
• The session is preconfigured: headed Chromium at /usr/local/bin/workspace-chrome on virtual display, persistent profile at /workspace/.browser-profile (cookies/logins survive across commands and tasks), session name workspace.
• Dashboard on port 4848 for human observation (signed URL via AgentBox Manager /sandboxes/<id>/browser-url).
• Local apps: browse http://127.0.0.1:<port> from inside the container, never the public preview URL.
• Never install Playwright or browser binaries — everything is preinstalled.

Acting On Pages

agent-browser fill @e3 "user@example.com"     # clear + type
agent-browser type @e4 "extra text"           # type without clearing
agent-browser press Enter
agent-browser select @e5 "Option A"
agent-browser check @e6 / uncheck @e6
agent-browser upload @e7 ./file.pdf
agent-browser scroll down 500 ; agent-browser scrollintoview @e9
agent-browser click @e8 --new-tab

No snapshot handy? Semantic locators work without one:

agent-browser find text "Sign In" click
agent-browser find role button click --name "Submit"
agent-browser find label "Email" fill "user@test.com"

Raw CSS selectors (agent-browser click "#submit") are the last resort.

Waiting (pick the right one)

Reading And Extracting

agent-browser get text @e5 ; agent-browser get attr @e10 href
agent-browser get url ; agent-browser get title
agent-browser --max-output 500000 get html html > page.html   # big output needs --max-output
agent-browser screenshot shot.png ; agent-browser screenshot --full full.png
agent-browser screenshot --annotate map.png                   # numbered labels keyed to @eN refs

# Arbitrary JS — heredoc avoids quote-escaping hell
cat <<'EOF' | agent-browser eval --stdin
Array.from(document.querySelectorAll("table tbody tr")).map(r => ({
  name: r.cells[0].innerText, price: r.cells[1].innerText,
}))
EOF

Save pages for later reading/citation (markdown via Readability+Turndown; pdf/jpeg/png direct):

save-webpage https://example.com/article --formats markdown,pdf --out research

Recipes

Login + persist. Fill the form via refs, wait --url "**/dashboard", done — the persistent profile keeps you logged in for later commands. For repeatable logins without secrets in shell history:

agent-browser auth save my-app --url https://app.example.com/login \
  --username user@example.com --password-stdin     # then: agent-browser auth login my-app
# Or save/reuse cookie state explicitly:
agent-browser state save ./auth.json
agent-browser --state ./auth.json open https://app.example.com

Tabs. agent-browser tab (list), tab new <url>, tab 2, tab close 2. Refs are per-page — re-snapshot after switching.

Parallel isolated sessions. agent-browser --session user-a open ... — own cookies, tabs, refs per session.

Dialogs and iframes. agent-browser dialog accept|dismiss; iframes are auto-inlined in snapshots (refs work through them), or agent-browser frame @e3 / frame main to switch context explicitly.

Local app debugging. curl -fsS http://127.0.0.1:<port> first → start-browser http://127.0.0.1:<port> → reproduce → screenshot + console/network logs — don't stop at the visual failure.

Test a pod app (authenticated)

To exercise a Lemma app in the agent browser as the current agent/user, let the CLI open it authenticated instead of wiring tokens by hand:

lemma apps open support-app                          # a DEPLOYED app, by slug — resolves its served URL + injects the bearer
lemma apps open --url http://localhost:5173 --no-auth # a `npm run dev` app — it already self-authenticates via the dev token

lemma apps open <slug> registers the current access token as an Authorization: Bearer header scoped to the API origin (so the app's cross-origin API calls authenticate), then opens the app — no login UI. A local dev server (npm run dev) seeds the token itself, so pass --url <dev-url> --no-auth. From there it's the normal core loop: snapshot -i → click/fill → re-snapshot, plus screenshot to capture the rendered UI.

See it with your own eyes. Take a screenshot, then use the view-image capability on that PNG to actually view the rendered app (layout, charts, broken styles, error overlays). view-image also reads pod and workspace files directly — a downloaded page render (lemma files child …/pages/page_0001.jpg), an uploaded image, or a local screenshot — so you can confirm a chart or document looks right without a browser. (App design, deploy, and test details: lemma-builder/references/apps.md.)

Troubleshooting

• Element missing from snapshot → scroll it into view, wait for it, or dismiss the overlay covering it; then re-snapshot.
• Click does nothing → a modal/banner is intercepting; find and dismiss it.
• Fill ignored by custom inputs → agent-browser keyboard inserttext "text" bypasses key events.
• CDP/connection errors or weird state → agent-browser doctor --fix.
• More guides ship with the CLI: agent-browser skills list, agent-browser skills get <name>; references/agent-browser-core.md has the full core reference.

See also

• Full core reference (snapshot/ref model, every command) → references/agent-browser-core.md
• Build, deploy, and design a pod app → lemma-builder/references/apps.md
• Operate the pod (open apps, mint file URLs) → the lemma-user skill
• Inline live views over pod data → the lemma-widget skill