agent-browser

name	agent-browser
description	Drive a real browser to validate web apps end-to-end. Use when you need to navigate to a URL, interact with UI elements, fill forms, click buttons, take screenshots, assert page state, or run a full end-to-end validation flow for a web feature. The primary tool for automated browser-based E2E testing in the AI Layer.

agent-browser — Automated E2E Browser Testing

agent-browser is a CLI-driven browser automation tool (by Vercel Labs) that the AI agent controls directly via shell commands. No Playwright/Puppeteer code required — the agent issues commands, reads the structured output, and makes decisions.

Installation

# Install globally
npm install -g agent-browser

# Download browser engine (Chrome for Testing)
agent-browser install

Platform notes:

• macOS/Linux: Works natively after install.
• Windows: Has a known issue with Unix domain sockets. Use WSL as a workaround, or run from a Linux container.
• Docker/CI: Install in the image with the two commands above.

Core workflow

1. Navigate: agent-browser open <url>
2. Snapshot: agent-browser snapshot -i — returns interactive elements tagged with refs (@e1, @e2, …)
3. Interact using those refs
4. Re-snapshot after navigation or significant DOM changes
5. Assert state with get, is, or wait commands
6. Screenshot for evidence: agent-browser screenshot path.png

E2E Testing Protocol

When using agent-browser to validate a feature end-to-end:

1. Start the dev/preview server if it is not already running (check CLAUDE.md for the project's start command).
2. Navigate to the feature's entry URL.
3. Snapshot to discover the interactive elements.
4. Exercise the happy path — fill inputs, click buttons, submit forms, assert success state.
5. Exercise key error/edge paths — missing required fields, invalid input, auth-required pages.
6. Screenshot at each key moment — before interaction, after success, after error. Save to screenshots/ with descriptive names.
7. Check console errors with agent-browser errors at the end of the session.
8. Close the browser: agent-browser close.

Commands

Navigation

agent-browser open <url>      # Navigate to URL
agent-browser back            # Go back
agent-browser forward         # Go forward
agent-browser reload          # Reload page
agent-browser close           # Close browser

Snapshot (page analysis)

agent-browser snapshot            # Full accessibility tree
agent-browser snapshot -i         # Interactive elements only (recommended for testing)
agent-browser snapshot -c         # Compact output
agent-browser snapshot -d 3       # Limit depth to 3
agent-browser snapshot -s "#main" # Scope to CSS selector

Interactions (use @refs from snapshot)

agent-browser click @e1           # Click
agent-browser dblclick @e1        # Double-click
agent-browser focus @e1           # Focus element
agent-browser fill @e2 "text"     # Clear and type
agent-browser type @e2 "text"     # Type without clearing
agent-browser press Enter         # Press key
agent-browser press Control+a     # Key combination
agent-browser hover @e1           # Hover
agent-browser check @e1           # Check checkbox
agent-browser uncheck @e1         # Uncheck checkbox
agent-browser select @e1 "value"  # Select dropdown option
agent-browser scroll down 500     # Scroll page
agent-browser scrollintoview @e1  # Scroll element into view
agent-browser drag @e1 @e2        # Drag and drop
agent-browser upload @e1 file.pdf # Upload file

Get information / Assert state

agent-browser get text @e1        # Get element text
agent-browser get html @e1        # Get innerHTML
agent-browser get value @e1       # Get input value
agent-browser get attr @e1 href   # Get attribute
agent-browser get title           # Get page title
agent-browser get url             # Get current URL
agent-browser get count ".item"   # Count matching elements
agent-browser get box @e1         # Get bounding box

agent-browser is visible @e1      # Check if visible
agent-browser is enabled @e1      # Check if enabled
agent-browser is checked @e1      # Check if checked

Wait / Synchronization

agent-browser wait @e1                     # Wait for element to appear
agent-browser wait 2000                    # Wait milliseconds
agent-browser wait --text "Success"        # Wait for text to appear
agent-browser wait --url "**/dashboard"    # Wait for URL pattern
agent-browser wait --load networkidle      # Wait for network idle
agent-browser wait --fn "window.ready"    # Wait for JS condition

Screenshots & PDF

agent-browser screenshot                  # Screenshot to stdout
agent-browser screenshot screenshots/step1.png  # Save to file
agent-browser screenshot --full           # Full page screenshot
agent-browser pdf output.pdf             # Save page as PDF

Semantic locators (alternative to @refs)

agent-browser find role button click --name "Submit"
agent-browser find text "Sign In" click
agent-browser find label "Email" fill "[e-mail]"
agent-browser find first ".item" click
agent-browser find nth 2 "a" text

Debugging

agent-browser open <url> --headed    # Show browser window (useful locally)
agent-browser console                # View console messages
agent-browser errors                 # View page errors (run at end of test)
agent-browser highlight @e1          # Highlight element
agent-browser trace start            # Start recording trace
agent-browser trace stop trace.zip   # Stop and save trace
agent-browser record start ./demo.webm  # Record video session
agent-browser record stop               # Stop and save video

JavaScript

agent-browser eval "document.title"       # Run JavaScript and get result

Network (mocking/interception)

agent-browser network route <url>              # Intercept requests
agent-browser network route <url> --abort      # Block requests
agent-browser network route <url> --body '{}'  # Mock response body
agent-browser network unroute [url]            # Remove routes
agent-browser network requests                 # View tracked requests

Sessions (parallel browsers)

agent-browser --session test1 open site-a.com
agent-browser --session test2 open site-b.com
agent-browser session list

Saved auth state

# Login once, save state
agent-browser open https://app.example.com/login
agent-browser fill @e1 "username"
agent-browser fill @e2 "password"
agent-browser click @e3
agent-browser wait --url "**/dashboard"
agent-browser state save auth.json

# Reuse in subsequent test runs
agent-browser state load auth.json
agent-browser open https://app.example.com/dashboard

JSON output (for programmatic checks)

agent-browser snapshot -i --json
agent-browser get text @e1 --json

Example: End-to-end form submission test

# 1. Start
agent-browser open http://localhost:3000/signup

# 2. Discover
agent-browser snapshot -i
# → textbox "Email" [ref=e1], textbox "Password" [ref=e2], button "Create account" [ref=e3]

# 3. Happy path
agent-browser fill @e1 "[e-mail]"
agent-browser fill @e2 "SecurePass123!"
agent-browser screenshot screenshots/before-submit.png
agent-browser click @e3
agent-browser wait --url "**/dashboard"
agent-browser screenshot screenshots/after-signup.png
agent-browser get title
# → Expected: "Dashboard"

# 4. Error path — submit empty form
agent-browser open http://localhost:3000/signup
agent-browser snapshot -i
agent-browser click @e3          # Submit without filling in
agent-browser wait --text "required"
agent-browser screenshot screenshots/validation-errors.png

# 5. Check for console errors
agent-browser errors

# 6. Clean up
agent-browser close

Integration with plan-feature validation levels

When a plan's Level 5 (E2E / Browser Automation) is reached:

1. Confirm the dev server is running.
2. Follow the E2E Testing Protocol above.
3. Save screenshots to screenshots/<ticket-id>-<description>.png.
4. Paste the screenshot paths and a pass/fail summary into the plan's completion checklist.

Notes

• Always re-snapshot after a navigation or significant DOM mutation — refs are invalidated.
• Prefer agent-browser wait --load networkidle after form submissions before asserting state.
• For CI environments, omit --headed; for local debugging, add it so you can watch the browser.
• agent-browser errors at the end of a session catches JS exceptions that wouldn't otherwise surface.