A browser your agents use and you can watch
An agent needs a browser to test the UI, and you need to see what it’s doing. AgentVera runs a separate-profile Chrome that agents use, and you watch it live on the Browser page.
How it works
Agents in “AgentVera browser” mode drive it through Playwright MCP (--cdp-endpoint). On the Browser page, tabs stream live over CDP screencast.
- Click, scroll, type, address bar, back and forward.
- Screenshots, DevTools and the console log.
- For Claude there’s also a “Claude in Chrome” (--chrome) option.
- Links from agents open in the in-app browser; ⌥-click opens the system browser.
Together with ports
The Ports tab in the right panel shows dev servers your agents started; open any port in the AgentVera browser in one click.
Questions
Does it touch my own Chrome profile?
No. The AgentVera browser uses a separate profile.
Which agents can use it?
Claude Code, Codex and CLI agents that take MCP. You pick it in the Browser field when creating the agent.
Can I see what the agent is doing?
Yes. Tabs stream live on the Browser page, and you can step in and click whenever you like.
Related features
Related terms:MCP (Model Context Protocol)Visual regression testing
Bring your agents to one desk.
Start on the free plan. Download it, add a project folder and open your first agent.
Posts on this topic
Blog- October 28, 2026Claude Code MCP setup: add, scope and manage serversHow to add MCP servers to Claude Code: claude mcp add for stdio and HTTP, scopes, .mcp.json, /mcp and OAuth, Claude Desktop import, timeouts and output limits.
- October 27, 2026How to debug code with AI: a practical workflowDebug code with AI the reliable way: reproduce first, hand the agent the failing test and logs, bisect with git, find the root cause, then lock the fix in a test.
- October 27, 2026Optimize code performance with AI: measure firstOptimize code performance with AI agents: profile first, give the agent real profiles, fix N+1 queries and bundle bloat, and prove each win with a benchmark.