Loongphy RSS

Codex Auth v0.2 Released: API Usage and Automatic Account Switching

Tue, 17 Mar 2026 00:00:00 GMT

I'm thrilled to announce the release of the Codex Auth v0.2 beta! 🎉

In this version, we've completely refactored the underlying architecture to solve the issue where a single regular account might be bound to multiple different Team workspaces. At the same time, to address the unreliable local session file usage tracking from the previous version, we've introduced a brand-new API mode. Furthermore, we've added an experimental background service for automatic account switching.

⚠️ STRONGLY RECOMMENDED: Please backup before upgrading

Before upgrading, please make sure to back up your current ~/.codex/auth.json file. If you have previously added multiple accounts using codex-auth, please also back up your entire ~/.codex/accounts directory just in case.

✅ Upgrade and Experience Now

v0.2 is still in beta. We only recommend enabling the API usage query after updating; as for the background account switching feature, we suggest waiting for continuous iterative updates.

Install the beta version

npm install -g @loongphy/codex-auth@next

Enable API usage query

codex-auth config api enable

📖 About Codex Auth

If you're not yet familiar with this project, Codex Auth is a local command-line tool designed specifically for managing and seamlessly switching between multiple Codex accounts. Its biggest highlight is its zero-intrusiveness: you don't need to modify any of Codex's configuration files, nor do you need to deploy any extra proxy services. You can easily manage and switch accounts with just simple command-line operations. Repository: https://github.com/Loongphy/codex-auth

Codex Auth is perfectly compatible with the three mainstream Codex usage scenarios:

Native Codex-CLI command-line invocation (when using the official codex-cli, you need to exit and restart after switching accounts, or use the modified version codext). For more details, see Building a Smoother Codex than the Official One: Automated Skills Sync with Upstream.
Independently running Codex App
Built-in invocation via the VSCode Extension

Here are the key highlights of this update.

💡 Core Features

1. API Usage Query

Initially, we had no intention of introducing any API calls to this fully localized tool, to avoid the risk of being monitored by OpenAI and causing account bans. However, tracking usage via local files (rollout-*.jsonl) has recently become highly unreliable, often resulting in rate_limits: null situations (for details, see OpenAI Codex Issue #14880).

For accuracy, we referred to the quota retrieval method of Codex-CLI and introduced support for real-time usage via API!

[!CAUTION] Please note: Enabling the API method to retrieve usage may violate OpenAI's terms of service and lead to account bans. Please use it with caution and at your own risk.

How to enable: This feature is disabled by default. You can manually turn it on by executing codex-auth config api enable to enjoy precise quota monitoring.
Future plans: This mode is planned to be the default usage synchronization method in v0.3.

2. Smart Background Auto-Switching 🧪 Experimental Feature

In high-frequency daily use cases, manually switching accounts after entirely depleting a quota often disrupts your state of flow. To solve this problem, v0.2 introduces a brand-new background daemon mechanism.

This process can continuously monitor your current usage status in the background. When your remaining quota drops below a set threshold, it will automatically switch to the most optimal available account for you.

[!TIP] The background auto-switching feature is currently still Experimental. At this stage, we have mainly conducted thorough stability testing in the Linux (WSL2) environment; for macOS and Windows platforms, we have merely implemented and retained compatibility support for now.

The default setting triggers a switch when the remaining quota drops to 10% within a 5-hour window or 5% for the weekly quota.

Common Commands:

# Enable
codex-auth config auto enable

# Disable background switching (default)
codex-auth config auto disable

# Customize warning thresholds
codex-auth config auto --5h 15 --weekly 10

3. A Boon for Multi-Subscription Users: Account Grouping

This is a fundamental architectural upgrade. Previously, we used email addresses as the primary key to store account information, but this could not satisfy complex scenarios like "a single email under multiple workspaces or different subscription plans (e.g., Team, Free/Plus/Pro coexisting)."

In v0.2, when you run list or switch, the program will automatically group and collapse multiple subscription plans under the same email alias. This makes it much more visually intuitive and clear, while still retaining the flexibility of independent switching.

Other New Commands

We have also greatly enhanced the routine maintenance commands, ensuring your environment remains stable and controllable under various extreme edge cases:

codex-auth status: View the background service status, auto-switch toggle, and API usage mode with one click.
codex-auth daemon --watch|--once: Run in the foreground or test the background switching logic.
codex-auth clean: A general cleanup command (not recommended for execution), used to remove out-of-date authorization snapshots and old backup files.
codex-auth import --purge: An exclusive command for disaster recovery. By utilizing existing authorization snapshot files, you can rebuild the registry from scratch. If things don't work correctly after an update, you can try using this command.

📦 For Contributors: Introducing PR Preview Deployments

To speed up the review and testing of community contributions, we have introduced pkg.pr.new support in our CI workflow.

Now, every Pull Request will automatically build and publish platform-specific preview packages, whilst still retaining the smart feature of parsing on-demand dependencies based on OS/CPU. You can even intuitively see the preview build source by using --version: codex-auth 0.1.5 (preview pr-6 b6bfcf5)

Want to quickly help test a PR? Just execute something like this:

npx https://pkg.pr.new/@loongphy/codex-auth@6

Want to test a build from a specific commit in that PR? Use something like:

npx https://pkg.pr.new/@loongphy/codex-auth@e32a1

If you encounter any issues while using it, please feel free to join us at our GitHub Repository to communicate or submit an Issue!

OpenAI Codex App for Windows: A Build-It-Yourself Tutorial (With Installer)

Wed, 04 Feb 2026 00:00:00 GMT

OpenAI has a habit of releasing macOS-only products first. Fortunately, this time around, Codex is built with Electron (a framework that lets developers build desktop apps using web technologies) for its UI, while relying on Codex CLI under the hood. That opens up the possibility of reverse-engineering and rebuilding a Windows version.

Pre-Built Installer

This is just for trying it out — a proof of concept, not a maintained release.

Download here: https://github.com/Loongphy/codext/releases/tag/rust-v0.94.0
Alternatively, check out CodexDesktop-Rebuild, which inspired this article. Note that it uses a modified version of Codex CLI, though I'm not sure what they changed.

I still mainly use Codex CLI directly — the app is rarely my go-to. After all, the UI is just a wrapper around Codex CLI, and I'm perfectly happy with my current workflow; I don't need a GUI. I also can't give up WSL2 (Windows Subsystem for Linux — a feature that lets you run a real Linux environment directly inside Windows). The Windows app can only access WSL files through cross-filesystem paths like \\wsl.localhost\Debian\root, which makes file operations painfully slow.

Feel free to drop questions in the comments. Before the official release, I'll keep updating alongside Codex CLI releases.

As the saying goes — teach a man to fish. I hope that after reading this article, you'll be able to build and fix things on your own 🫡.

Downloading the macOS Installer

You'll need to trick Chrome into pretending it's running on macOS — by spoofing the User Agent (the string your browser sends to websites to identify itself) and Client Hints (a modern mechanism browsers use to share device info). Then open https://openai.com/codex/ to download the codex.dmg installer (.dmg is the standard macOS installer format):

Open DevTools: press F12 or Ctrl+Shift+I
Open the command menu: Ctrl+Shift+P, type Network conditions, and select Show Network conditions (a panel will appear at the bottom of DevTools).
In the User agent section, uncheck Use browser default, then:
- Select a macOS / Safari option from the dropdown (if available)
(This is crucial) Expand the User agent client hints dropdown and change the platform field to macOS, then click Update. Many websites nowadays rely more heavily on Client Hints than the traditional UA string, so this step is what actually makes the spoofing work.
Refresh the page.

Building the Codex Windows App

I much prefer developing inside WSL2 with CLI-based agents. The Windows terminal experience is an absolute nightmare — LLMs tend to make more mistakes there too, and accidental file deletions are all too common.

Create an empty folder, copy the downloaded .dmg file into it, open this directory in WSL2, run codex, type /Approvals to grant Full Access permissions, and let's go!

Initial Prompt

Based on the current state of my repository, help me repackage a "macOS-packaged Electron application" into a version that can be built and packaged on Windows.

## Project Goal
- I only have a macOS installer (a post-build Electron .dmg). I want to rebuild it into a working Windows app.
- I'm not trying to recover source code or recompile the frontend. The goal is "repackaging": take the already-packaged output as input and rebuild it into a runnable Windows app.
- The backend CLI is provided by the Codex CLI process: on Windows we need `codex.exe`, obtained via the `@openai/codex` package which provides pre-compiled binaries for each platform.

## Deliverable

- A Windows installer

## Notes

- We're in a WSL2 environment. For Windows-specific dependency packaging, you'll need to invoke `pwsh.exe`.

Being an Observer

AI users tend to let the tool fix every error for them, rarely thinking independently. Over time, this weakens their ability to spot and fix bugs on their own — which is precisely the skill you need most when supervising AI output.

I've seen plenty of people doing Vibe Coding (writing code by feel, leaning heavily on AI) who only look at the final result and blame the agent, without ever examining the process to see where things went off track.

Here are a few issues I observed while watching Codex CLI work:

prefer uv not python

During execution, I noticed it was using Python. Since I didn't want it leaving junk all over my system by installing random dependencies, I added a prefer uv not python instruction mid-run.

That said, it still went ahead and installed Python directly rather than using uv, so I'd recommend putting this kind of requirement in your initial prompt or your AGENTS.md file (a configuration file that Codex CLI treats as its highest-priority instruction set).

Codex CLI treats instructions in AGENTS.md as top priority and follows them strictly — it rarely overlooks them.

Install Visual Studio Community

The packaging process uses Visual Studio Tools, so you may need to install Visual Studio Community first. You'll need the same setup as Electron development requires: install the Desktop development with C++ workload.

The Result: One Run

With just the initial prompt and one mid-run instruction (prefer uv not python), the whole thing finished in 41 minutes.

It did run successfully, though there were still a few bugs — the window wasn't properly resizable, and selecting a project triggered an Oops, an error has occurred message.

All of these were fixed with a single follow-up prompt. Now, let's play with it!

All hail Codex CLI!

An Alternative Build Path

Maybe you don't want to build locally or deal with installing a bunch of dependencies. You can try running Codex CLI inside GitHub Actions (GitHub's built-in automation platform) — set it up to build in a Windows environment using codex "your instructions", then upload the packaged output as an Artifact so you can download it directly.

You might need to tweak the prompt to make it keep retrying until the build succeeds, since Codex CLI tends to ask questions rather than just powering through autonomously.

Looking Ahead

Building a proper Windows version isn't that simple. I don't expect them to use WinUI 3 — even Microsoft itself is stuck in a technical quagmire on that front. But I do hope that every developer building programming tools will give some thought to WSL2 compatibility and deliver a native Linux development experience on par with what VS Code offers.

Acknowledgments

Inspired by https://github.com/Haleclipse/CodexDesktop-Rebuild

How to Fix Google Antigravity Stuck on "Setting Up Your Account"

Wed, 19 Nov 2025 00:00:00 GMT

Non-US regions can also work, but you can't be in a region where Gemini isn't supported (like Hong Kong or Macau). If you want to be among the first to try new AI models and products, the US is still your best bet.

Today (2025-11-19), Google released the Gemini 3 Pro model along with a brand-new AI code editor called Antigravity. The tool was built by former Windsurf core team members on top of Windsurf's codebase — you can still spot quite a bit of legacy code floating around.

In July 2025, Google hired Windsurf CEO Varun Mohan, co-founder Douglas Chen, and several researchers to join DeepMind, while also securing a non-exclusive license to parts of Windsurf's technology. The deal totaled $2.4 billion, primarily for talent acquisition and technology licensing — not a full acquisition of the company.

Previously, OpenAI had attempted to acquire Windsurf for $3 billion, but the deal fell through.

After the launch, I kept getting stuck on the "Setting Up Your Account" screen right after logging in. I even tried using Mihomo (a popular proxy client) in TUN mode (a system-wide proxy mode), but that didn't help either. After digging through community discussions, I found out the issue was tied to the account's associated region.

Visit https://policies.google.com/terms — if it says "Hong Kong," congratulations, you won't be able to use Antigravity.

Antigravity restricts access for regions where Gemini isn't supported. But here's the catch: Gemini only checks your IP address, while Antigravity goes a step further — it checks both your IP and your Google account's associated region.

With that understood, the solution is straightforward:

Register a new Google account from a US network environment
Transfer your existing Google account to a US region

Transferring Your Google Account Region

Submit a region change request at: https://policies.google.com/country-association-form
Select the United States, pick any state, choose Other Reason, and explain that you need access to Google AI products. For example:
```
I need to use Google Gemini and Antigravity, which requires a US-based account context.
```

In my experience, the approval email came through in about 30 minutes. A word of advice: don't pick any other reason for the transfer — those tend to get rejected.

Summary: Getting Antigravity to Work

US proxy node + TUN mode
A US-based Google account — verify at https://policies.google.com/terms. If it doesn't show the US, follow the region transfer steps above.

From what others in the community have shared, non-US regions like Japan also work. Personally, I just went with the US for simplicity.

Gemini in Chrome

Unlike Antigravity, it hasn't been confirmed yet whether non-US accounts can use this feature. If you can't enable it, consider transferring your account region to the US.

With a US Google account set up, you can also enable Gemini in Chrome — giving your browser a built-in Gemini AI assistant. It's quite handy for things like quickly summarizing a web page.

Change your browser language to: English (United States)
Go to chrome://flags/ and enable Glic and Tabstrip Combo Button.
Restart your browser — you'll see a new Gemini icon in the top-right corner.

I had to switch Chrome to English (United States) before Gemini showed up. Some people say it also works with Chinese — go figure.