Connect ChatGPT desktop and Codex
One button adds Premation to the setup ChatGPT desktop, the Codex CLI and the Codex IDE extension share.
- 1
Open Connections in Premation
On Premation's home screen, choose Connections in the sidebar. Inside the editor the same cards are under Settings ▸ AI Engine.
- 2
Click Connect ChatGPT
It is on the Cursor, Antigravity and ChatGPT card. Premation adds its own entry to ~/.codex/config.toml, writes a backup first, and leaves the rest of the file alone.
- 3
Restart ChatGPT desktop or Codex
They read the file when they start. On Premation's card, the ChatGPT row now reads Ready.
- 4
Ask for something
Start with "In Premation". Premation gives ChatGPT up to 10 minutes per step, so a long build is not cut off half way.
Try In Premation, make a vertical 15-second promo for a summer sale with three scenes.
How you know it worked
Setup details on the ChatGPT row shows the file Premation wrote and the entry it added.
If something is wrong
- The assistant says Premation is not open
- Open Premation and a project, then ask again. The assistant works through the running app.
- The assistant says to sign in
- Sign in to Premation, then ask again.
- The row says Paused
- The app is set up, but Premation is not accepting assistants right now. Click Turn on.
- You would rather not have the file edited
- Open Setup details on the ChatGPT row and run the codex mcp add command shown there instead.
- The row says Needs a look
- Premation could not read config.toml and will not change it. Fix the file, or add Premation by hand with Copy snippet.
- The row says Needs repair
- Click Repair, then restart ChatGPT or Codex. Premation moved since it was connected, usually after an update.
Turning it off
Click Disconnect on the ChatGPT row. Only Premation's entry is removed from the file.