Someone built an MCP plugin at the Codex Hackathon that connects ChatGPT and Codex.


The problem it solves is very specific: GPT 5.5 Pro is currently the most planned and researched model, but it can't be used inside Codex.
Codex excels at execution, but its planning ability is not as good as 5.5 Pro.
His approach is to use MCP to expose the local development environment to ChatGPT, letting 5.5 Pro handle planning and task breakdown, then delegating execution to Codex.
It's like letting two AIs do what they're best at: one thinks, the other acts.
There's also an unexpected benefit: the rate limits for ChatGPT and Codex are calculated separately.
With this setup on the same account, the quota effectively doubles.
This idea is actually quite worth learning from.
Many people struggle with "which AI tool to use," but the real answer might be: use them all and let them work together.
The barriers between AI tools are being broken down by protocols like MCP, and in the future, it might not be about choosing the strongest one, but about assembling the most suitable combination.
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pinned