Browser actions
Read a page and its interactive elements, click or fill a field, and wait for the page to respond.
A command-line tool with a companion Skill that lets your agent operate Lexmount Cloud Browser, manage sessions, export files, and debug browser tasks.
Read a page and its interactive elements, click or fill a field, and wait for the page to respond.
Create and close sessions, list open tabs, and choose the tab each command acts on.
Save screenshots and PDFs, and retrieve files downloaded during a session.
Check your connection, run JavaScript on the page, and send Chrome DevTools Protocol commands.
--target-id. Before clicking, the CLI checks that the element is visible, enabled, and unobstructed.Send the prompt to your agent, then approve access when asked.
Once installed, try sending this to your agent
Use Lexmount Browser CLI to open https://www.selenium.dev/selenium/web/web-form.html, enter “Hello from Lexmount” in Text input, and submit the form. When “Received!” appears, save a screenshot of the result.
Rename the download to browser-cli and run it from its directory.
chmod +x browser-cli
./browser-cli auth login --connect-base-url https://browser.lexmount.com
./browser-cli doctorRename the download to browser-cli.exe and run it from its directory.
.\browser-cli.exe auth login --connect-base-url https://browser.lexmount.com
.\browser-cli.exe doctorRename the download to browser-cli and run it from its directory.
chmod +x browser-cli
./browser-cli auth login --device --connect-base-url https://browser.lexmount.com
./browser-cli doctorWhen doctor returns ready_for_browser_actions: true, the cloud browser is ready.
| Command | Purpose |
|---|---|
browser-cli auth login --client-name "Agent" | Authorize in the browser and save local credentials |
browser-cli auth login --device | Approve access for SSH or a headless terminal on another device |
browser-cli auth status | Check the current credentials |
browser-cli doctor | Check credentials and API readiness |
| Command | Purpose |
|---|---|
browser-cli session create | Create a cloud browser session |
browser-cli session targets --session-id SESSION_ID | List the pages in a session |
browser-cli session get --session-id SESSION_ID | Get session status and the viewing URL |
browser-cli context create | Create a reusable browser context |
browser-cli context list | List contexts and their usage status |
browser-cli session create --context-id CONTEXT_ID --context-mode read_write | Start a session with an existing context |
browser-cli session close --session-id SESSION_ID | Close a specific session |
| Command | Purpose |
|---|---|
browser-cli action open-url --session-id SESSION_ID --url https://example.com | Open a URL in the session |
browser-cli action snapshot --session-id SESSION_ID | Read page content and interactive elements |
browser-cli action fill --session-id SESSION_ID --selector "#query" --value "browser cli" | Fill a specific input field |
browser-cli action click --session-id SESSION_ID --selector "#search" | Check the target and send a mouse click |
browser-cli action wait-text --session-id SESSION_ID --text "Saved" | Wait for matching text on the page |
browser-cli action wait-selector --session-id SESSION_ID --selector "#results" | Wait for an element to appear |
browser-cli action snapshot --session-id SESSION_ID --target-id PAGE_ID | Choose a specific browser tab |
| Command | Purpose |
|---|---|
browser-cli action screenshot --session-id SESSION_ID --path page.png --full-page | Save a full-page screenshot |
browser-cli action pdf --session-id SESSION_ID --path page.pdf --print-background | Export the page as a PDF |
browser-cli session downloads list --session-id SESSION_ID | List files downloaded during the session |
browser-cli session downloads get --session-id SESSION_ID --download-id DOWNLOAD_ID --output download.bin | Save a session download to your machine |
browser-cli action eval --session-id SESSION_ID --expression "document.title" | Run JavaScript on the page |
browser-cli tunnel open | Open and maintain a local network tunnel |
browser-cli session create --local-proxy-tunnel-id TUNNEL_ID | Route a session through a local tunnel |
browser-cli tunnel status --tunnel-id TUNNEL_ID | Inspect the network tunnel status |
browser-cli tunnel close --tunnel-id TUNNEL_ID | Close a specific network tunnel |
browser-cli action raw --session-id SESSION_ID --method Page.getLayoutMetrics --params-json "{}" | Call a Chrome DevTools Protocol method |
No. Native executables are available for Apple silicon Macs, Windows x64, and Linux x64. The official Skill’s setup script downloads a pinned release and verifies its checksum.
Both use the same CLI. The official Skill guides your agent through setup, authorization, inspecting pages, taking actions, and closing sessions. You can also run those commands yourself in a terminal.
Check the result of each command. A successful click means the input was sent. When waiting for text, check matched and timed_out. Use a snapshot, screenshot, or a check for the expected page state to verify the outcome.
Run session targets and identify the tab by its URL and title. Pass its --target-id to subsequent action commands. The CLI reports an error if that target disappears.
Create a persistent Context, then pass its ID and access mode when creating a session. Close the read-write session normally to save its state. Use that same Context for your next session.
The CLI source is available under the MIT license. Lexmount Cloud Browser is a hosted service. Usage is billed according to your plan. See pricing for allowances and rates.