Integrations / AI agent clients / Codex CLI

A browser for Codex CLI.

Codex CLI works in your terminal and repository. Add Browz over MCP and it can check the running result of its changes in a real browser, each task in an isolated profile with its own cookies, so it never sees your own sessions.

Terminalcodex mcp add browz -- /Applications/Browz.app/Contents/Resources/browz-mcp

Setup

Set it up.

  1. Install Browz and open it

    Download Browz; the MCP server ships inside it.

  2. Add Browz

    The server goes into ~/.codex/config.toml, for every project.

    Terminalcodex mcp add browz -- /Applications/Browz.app/Contents/Resources/browz-mcp
  3. Or edit config.toml

    The same entry by hand. A trusted project can also have its own .codex/config.toml.

    ~/.codex/config.toml[mcp_servers.browz] command = "/Applications/Browz.app/Contents/Resources/browz-mcp"
  4. Check it is connected

    Run codex mcp list, or /mcp inside Codex, and look for Browz with its tools.

    Terminalcodex mcp list
  5. Remote (beta)

    Codex reads the bearer token from an environment variable you name, so the token never goes in the file. Create an agent token on your account page and put it in BROWZ_AGENT_TOKEN, or leave the line out and run codex mcp login browz to sign in.

    ~/.codex/config.toml, remote[mcp_servers.browz] url = "https://mcp.gobrowz.com/mcp" bearer_token_env_var = "BROWZ_AGENT_TOKEN"

Config format from Codex MCP docs.

Example

Check a change end to end.

After Codex changes a page, let it look at the result in a browser before it tells you it is done.

PromptStart the dev server, then use the browz MCP tools: open a throwaway session for "header check", go to http://localhost:8080, take a screenshot, check that the header shows the new logo, and close the session. Stop the dev server when you are done.

Good to know

Notes and tips.

  • The snippets use the macOS path of the server inside Browz. On Windows it is %LOCALAPPDATA%\Programs\Browz\browz-mcp.exe, on Linux /usr/lib/browz/browz-mcp; or use npx -y @browz/mcp, which finds it.
  • The remote server at https://mcp.gobrowz.com/mcp is in free public beta: it needs the Browz app open with Remote control on, on a computer with a hardware security chip, and offers the profile, tab and page tools. See what it needs and what it offers.
  • Browz has to be installed and open on the same computer: the local MCP server drives the app’s own profiles through its local API, and reads the app’s credentials itself, so no token goes in the config.
  • A person completes sign-ins, 2FA codes and CAPTCHAs. The agent can pause with browz_request_human and hand you the profile’s window, then carry on in the same session. Browz never solves or bypasses CAPTCHAs.
  • Use agents only where each site’s terms allow automated use, on accounts you are entitled to use. Our acceptable use policy applies to agents the same as to people.
  • The MCP server is on every plan, Free included (5 profiles). How many agent sessions can be open at once depends on the plan: see pricing.

FAQ

Codex CLI and Browz.

Is this the same as the OpenAI Agents SDK setup?

No. Codex CLI is a coding agent that reads config.toml; the Agents SDK is a library you write code with. Both reach the same Browz MCP server. See Browz for OpenAI agents for the SDK.

Does Codex need my Browz token?

Not for the local server: it reads the app’s own credentials. Only the remote server needs a sign-in or an agent token.

Can I give Codex fewer permissions?

Yes. Create a local agent key limited to the profiles and sites it needs and pass it to the server as BROWZ_LOCAL_AGENT_TOKEN in the env table of its config.

Get started

Download Browz. Every task, its own browser.

Then, on macOS or Linux, put the MCP server on your PATHcurl -fsSL https://gobrowz.com/agents/install.sh | sh

For Mac, Windows and Linux. Free for 5 profiles. Paid plans from $4.99 a month, with a 14-day refund on your first payment.