Skip to content
Codex Chrome extension launch | My AI Guide

Codex Chrome extension launch

By Harsh Desai
Share

TL;DR

The new Codex Chrome extension enables parallel interaction with apps and websites in your browser. It operates in the background, allowing you to maintain control over which websites Codex can access.

What changed

OpenAI released the Codex Chrome extension on May 18, 2026. The extension lets Codex interact with browser tabs and web apps in parallel while running in the background.

Users keep explicit control over allowed domains and can toggle access per site. No new pricing tier was announced. The feature appears under the existing Codex app subscription.

Why it matters

This move pushes Codex deeper into daily workflows that mix code, research, and web tasks. Solo builders gain an agent that can read dashboards, fill forms, and pull data without constant tab switching.

It directly competes with browser automation tools and lighter agents that lack deep model context. The bet is that persistent, permissioned web access will become table stakes for any serious coding agent by late 2026.

How to use it

Install the extension from the Chrome Web Store by searching for Codex. After installation, open the Codex app settings and grant site permissions only for domains you want the agent to touch.

Test with a simple prompt such as summarizing a Notion page or pulling data from a SaaS dashboard. The extension respects the same approval modes already configured in your Codex workspace.

Watch for

Confirm the bet if usage logs show frequent background actions on approved sites without permission prompts. The bet breaks if sites add stronger bot detection that forces constant manual approvals. Expect OpenAI to add selective DOM targeting or session recording next.

Harshs take

For a solo operator running client work through Notion and Zapier, the extension removes one layer of friction between research and output. The real cost is another always-on process watching your tabs, which adds both latency and a new attack surface you must monitor.

You trade some privacy and mental overhead for speed on repetitive web lookups. Most builders will over-grant access at first and regret it when an agent pulls the wrong data or hits rate limits.

Install the extension today on a secondary profile, limit it to three test domains, and measure actual time saved on your current projects before rolling it into production work.

by Harsh Desai

Source:developers.openai.com

About OpenAI Codex

View the full OpenAI Codex page →All OpenAI Codex updates

More from OpenAI Codex

  • App Update
    Codex CLI 0.129.0 release

    Codex CLI 0.129.0 adds modal Vim editing in the composer, redesigned TUI workflows for easier resuming and forking, enhanced plugin management with workspace sharing, and improved Linux sandbox reliability.

  • App Update
    Codex CLI 0.130.0 release

    Codex CLI 0.130.0 introduces plugin hook details, a new 'codex remote-control' command for headless app-server management, thread pagination, and improved Bedrock authentication using AWS login profiles.

  • Feature
    Expanded Auto-review documentation

    New documentation for Auto-review covers the reviewer lifecycle, trigger conditions, failure behavior, and configuration options. Updates also clarify how Auto-review interacts with the sandbox boundary.

Everything AI. One email.
Every Monday.

New tools. Model launches. Plugins. Repos. Tactics. The moves the sharpest builders are making right now, before everyone else.

No spam. Unsubscribe anytime.