miuuyy/codex-chatgpt-web

Use ChatGPT Web (including Pro) as a native model in Codex — with context, tools, streaming and images, without using Codex quota.

View on GitHub ↗Jump to charts ↓

Summary Information

Updated 59 minutes ago
Added to GitGenius on September 10th, 2026
Created on July 26th, 2026
Open Issues & Pull Requests: 53 (+1)
GitHub issues: Enabled
Number of forks: 597
Total Stargazers: 6,620 (+14)
Total Subscribers: 15 (+0)

Charts & Analytics

Fetching additional details & charts...

Issue Activity (beta)

Issue API getrepoissuespagesummary failed: 429 Rate limit exceeded. Please try again later.

Detailed Description

Codex ChatGPT Web is a bridge that integrates ChatGPT Web, including Pro tier, as a native model option within Codex while preserving Codex's task context, UI, and tool ecosystem without consuming Codex quota.

The tool solves the problem of accessing ChatGPT's capabilities within Codex's workflow by routing selected model tasks through a task-bound ChatGPT Temporary Chat. It maintains Codex's native task lifecycle, context management, and streaming while attaching images and exposing ChatGPT's reasoning and tool activity back into the same Codex interface. In full mode, MCP connects ChatGPT back to the tools of the active Codex task, allowing sequential messages to reuse a single Temporary Chat session that persists until a context boundary, at which point a checkpoint is written before starting a clean chat.

Developers should choose this tool if they want to leverage ChatGPT Web's capabilities without leaving Codex's native environment or consuming API quota. It suits projects where maintaining Codex's task UI, context lifecycle, and integrated tool access is important while switching between different model tiers. The tool runs as a cross-platform launcher on macOS, Windows, and Linux, handling sign-in, model setup, and health checks. Free and Go accounts access ChatGPT Web Luna, while accounts with reasoning selector access expose Instant, Medium, High, Extra High, and Pro tiers according to subscription level.

The project demonstrates active maintenance with continuous integration checks passing and comprehensive documentation including troubleshooting and security guidance. The codebase is written in TypeScript and includes release validation documentation covering end-to-end scenarios. The tool implements fail-closed behavior, producing explicit errors when models, tools, or ChatGPT UI elements change rather than silently degrading functionality.