The latest articles on DEV Community by Fluxenith (@fluxenith). https://dev.to/fluxenith https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4197937%2F95147f6c-07c3-402a-a42e-06cd48695547.png https://dev.to/fluxenith en Fluxenith Wed, 21 Oct 2026 10:02:10 +0000 https://dev.to/fluxenith/i-built-codexgpt-bridge-to-use-chatgpt-web-models-inside-codex-open-source-kfi https://dev.to/fluxenith/i-built-codexgpt-bridge-to-use-chatgpt-web-models-inside-codex-open-source-kfi <p>Hi DEV! </p> <p>I've been experimenting with a way to keep Codex's coding workflow while using the ChatGPT Web models available through my account.</p> <p>Switching between the two manually gets tedious: I have to move context across windows, paste responses back into the coding tool, and then check whether a suggested fix actually passes tests. So I built <strong>CodexGPT Bridge</strong>.</p> <p>It's an <strong>MIT-licensed, open-source desktop project</strong> for Windows, macOS, and Linux.</p> <p><strong>GitHub:</strong> <a href="https://github.com/Fluxenith/codexgpt-bridge" rel="noopener noreferrer">https://github.com/Fluxenith/codexgpt-bridge</a></p> <h2> What it does </h2> <p>The idea is to let each side handle what it is good at:</p> <ul> <li> <strong>ChatGPT Web</strong> provides model responses from a signed-in session.</li> <li> <strong>Codex</strong> keeps control of the local workspace, file changes, commands, tests, and tool approvals.</li> <li> <strong>CodexGPT Bridge</strong> connects the two so the task can move between reasoning and execution.</li> </ul> <p>I wanted this to feel like a normal coding session rather than a copy-and-paste conversation.</p> <p>The app also supports separate tasks, image input, progress updates, and cancellation, which matter once you move beyond a single prompt.</p> <h2> A real example </h2> <p>I recorded a <strong>46-second demo</strong> showing a bug being reproduced, a fix being applied, and the tests passing afterward. It's included in the repository README:</p> <p><a href="https://github.com/Fluxenith/codexgpt-bridge" rel="noopener noreferrer">See the demo and source code on GitHub</a></p> <p>For me, that end-to-end loop is the interesting part: the model can help think through the problem while Codex still performs and verifies the actual development work.</p> <h2> How to try it </h2> <ol> <li>Open the <a href="https://github.com/Fluxenith/codexgpt-bridge" rel="noopener noreferrer">GitHub repository</a> and follow the README setup instructions.</li> <li>Set up the desktop bridge and the required Codex/MCP integration and permissions.</li> <li>Sign in to your ChatGPT Web account.</li> <li>Start with a small coding task, review any requested tool approvals, and inspect the result.</li> </ol> <p>A few practical notes:</p> <ul> <li>Model availability and usage limits depend on <strong>your ChatGPT account</strong>.</li> <li>The full integration needs the documented MCP/plugin setup and permissions.</li> <li>The macOS builds are <strong>not yet notarized</strong>.</li> <li>This is an independent community project and is <strong>not affiliated with OpenAI</strong>.</li> </ul> <h2> I'd love feedback </h2> <p>The project is still evolving. If you try it, I'd especially like to hear about setup friction, longer-running tasks, and cases where the handoff between model reasoning and local tools feels awkward.</p> <p>Issues, suggestions, and contributions are welcome in the repository.</p> <p><strong>Would you use something like this in your coding workflow? What would you want it to handle better?</strong></p> showdev opensource ai devtools