Decide These Things Before Building Scenarios with the Make Plugin in ChatGPT

What can you do with the Make plugin in ChatGPT?

The official post describes a flow where you build, run, browse, and review the results of automations inside the conversation. You describe the result you want, and Make builds the scenario, connects the apps it needs, and sets the schedule or trigger. One example is an agent that watches new inquiries in an inbox and writes reply drafts, set up so a person approves each draft.

You can browse the scenarios you built by name, folder, and status, and review run results and history in a table with status indicators. If more tweaking is needed, open it in Make's visual editor and fix it there.

Can you install it now?

Check the status first. The official post says OpenAI temporarily disabled this plugin while it updates its automation tool support. The Make platform itself is not affected, and in the meantime the post suggests building through MCP from ChatGPT or Codex as an alternative. That is the status as of the date checked, so it is safer to check again in the OpenAI plugin directory before installing.

Per the official post, the installation steps are five.

1. Check that you have a Make account and a ChatGPT account.
2. Find Make in the plugin menu of ChatGPT or Codex and install it.
3. When prompted, sign in with your existing Make account.
4. Describe the result by writing the trigger, the data source, the task, and where to send the results.
5. Run the scenario once from the conversation and check the output.

What should you decide before building?

First decide which Make organization to build in. The official post says the scenario is created in the Make organization you signed in to. On a team with several organizations, it is easy to skip past confirming which one is selected.

Next is permissions. Installing the plugin does not lift each service's permissions or the workspace policy. Each app needs its own authentication, and in a managed workspace, the admin decides whether members can install it. Finally, write one sentence each for the input and the result. You need to be clear on what the trigger is, what data goes in, and where it goes out so you can properly compare the result of the first run.

Does the automation keep running if you close the conversation?

The official post explains that automations, app connections, stored data, permissions, and governance all stay in the Make organization. It says the plugin is a channel for ChatGPT to reach the Make organization, not a way of copying data out. Scenarios run in the Make cloud without ChatGPT being open, and Make handles scheduling, retries, and error handling.

So after a trial run, before turning on a schedule, check Make's run history and error notification settings. For actions that are hard to undo, such as sending email, it is better to have it stop at drafts and keep an approval step.

Source: Make official blog. Date checked: 2026-10-11.