Codex CLI can write and run code — but the moment it needs to log into GitHub, read an internal dashboard, or navigate a staging environment, it's stuck. VibeBrowser connects Codex to a pre-authenticated cloud browser over MCP. One command. No CDP plumbing. Markdown output that keeps context small.
npx @vibebrowser/mcp --remote YOUR_UUID. Codex CLI discovers the MCP tools immediately — navigate, snapshot, click, fill, scroll, new tab.
Codex writes the code, VibeBrowser opens the browser to test it. Build → deploy → verify in one agent loop without switching tools.
Transfer your logins once. Codex reaches GitHub, your CI dashboard, staging environments, and internal tooling — no re-auth, no CAPTCHA interruptions.
Codex gets clean structured content (~1–2 KB) instead of raw HTML (200–400 KB). Faster agent loops, lower token cost, no context overflow on multi-step tasks.
npx @vibebrowser/mcp --remote YOUR_UUID. Codex CLI discovers the MCP tools immediately — navigate, snapshot, click, fill, scroll, new tab.
Codex writes the code, VibeBrowser opens the browser to test it. Build → deploy → verify in one agent loop without switching tools.
Transfer your logins once. Codex reaches GitHub, your CI dashboard, staging environments, and internal tooling — no re-auth, no CAPTCHA interruptions.
Codex gets clean structured content (~1–2 KB) instead of raw HTML (200–400 KB). Faster agent loops, lower token cost, no context overflow on multi-step tasks.
The cloud browser runs 24/7. Codex can monitor builds, watch for failing tests, and check deploys overnight — no laptop required.
Passwords are never transmitted. VibeBrowser transfers session state — cookies and tokens — not credentials.
Ask Codex to open a failing GitHub Actions run and diagnose the error from the log page
Have Codex navigate to a staging URL, test a user flow, and file a GitHub issue if something breaks
Let Codex open an open PR, read the diff, and post a review comment via the GitHub UI
Ask Codex to check your Vercel deployment dashboard and confirm the latest deploy succeeded
Have Codex read an internal wiki page for context, then update the related code
Give Codex a list of npm packages to audit: browse each changelog and summarise breaking changes
Google Workspace, MCP access for other agents, reusable skills, and a secure secrets vault.
Gmail + Calendar actions built in for real workflows, not brittle APIs.
Expose your browser as an MCP server so other AI agents can drive it.
Securely connect remote agents to your logged-in browser without VPNs.
Create reusable skills and share them across your team or clients.
Internal vault with a password-fill tool that never exposes secrets to the LLM.
Works with Vibe AI, Anthropic Claude Max, GitHub Copilot, and BYOK providers.
Codex CLI is strong at writing and running code. VibeBrowser adds the browser layer: reading live web pages, clicking through UIs, testing deployed apps, and navigating authenticated portals. Together they cover the full build → test → verify loop.
Add VibeBrowser to your MCP config and run npx @vibebrowser/mcp --remote YOUR_UUID. Codex will discover the browser tools (navigate, snapshot, click, fill) automatically. Full guide at vibebrowser.app/mcp.
Yes for local mode (free tier). Run npx @vibebrowser/mcp without --remote and VibeBrowser connects to your local Chrome, which can reach localhost. Cloud mode uses a cloud IP, so it can't reach your local dev server directly.
Playwright needs page-specific selectors, handles authentication separately, and returns full DOM/HTML. VibeBrowser is model-first: pre-authenticated, markdown output, zero selector boilerplate. Codex describes what it wants to see — VibeBrowser handles the rest.
Yes. VibeBrowser is a standard MCP server — any OpenAI-model agent that routes tool calls through MCP can use it, including custom agents built on the OpenAI Agents SDK.