BuildGist

You’ve hit your usage limit in Codex: what to do next

Kayvan Zahiri ·

When Codex says You’ve hit your usage limit. and when to try again, your ChatGPT plan's Codex usage has run out. The files it already changed stay in the project folder. Wait for that time, buy more credits, or continue the task in Claude Code or Cursor.

The message, word for word

This is the message as Codex showed it on October 10, 2026, on a Pro plan:

You’ve hit your usage limit. Visit https://chatgpt.com/settings/usage to purchase more credits or try again at Oct 14th, 2026 9:19 AM.

The time is your computer's local time. A time alone, such as 9:14 PM, means later today. A date means another day, and a date several days out points to a weekly limit. Codex's pricing page, read on October 13, 2026, gives Plus a five-hour limit and Pro none, and says weekly limits may also apply.

The words before "or try again at" depend on your plan. On a Plus plan, the message also offers an upgrade. In September 2026 it read:

You've hit your usage limit. Upgrade to Pro (https://chatgpt.com/explore/pro), visit https://chatgpt.com/codex/settings/usage to purchase more credits or try again at 9:14 PM.

On a Business plan, it asks you to send a request to your admin instead. Without a reset time, the message ends with "try again later".

Your options, in order

  1. Wait for the time in the message. In a Codex session, /status shows the limits you have left.
  2. Buy credits. On Plus and Pro, the usage page the message links to is where you buy more, without changing your plan. On a Business plan, ask your admin.
  3. Continue the task in Claude Code or Cursor. If you already use one of them, you can keep going now. The rest of this guide is how.

Read what Codex already changed

In the project folder:

git status
git diff --stat
git diff
git log --oneline -5

You do not need to commit. git status also lists edits that were yours before the run, and new files stay untracked until added. Codex sometimes commits during a run, so read any commits from it in the log with git show. Read sign-in, permissions, deletion, payments, new dependencies and configuration first. The whole read is in how to review what Codex changed.

Write the task down

Claude Code and Cursor can read the files. They cannot read the Codex chat. Write a short note before you leave it. Use your own task. This sample uses the Google sign-in task from the BuildGist home page, where it is a real Claude Code run, shortened and edited.

Task: Add login with Google.
Done so far: the sign-in button, where Google sends you back, and the browser's signed-in
status. See the diff.
Not done: deleting a note does not check that the note belongs to the signed-in user.
Next step: add that check.
Stopped because Codex hit a usage limit. Do not redo the sign-in work.

Without that last line, the next agent can start the task over and edit files that were already right.

Open Claude Code or Cursor in the same folder

Start Claude Code in the project folder, or open Cursor on it and start a new chat. Paste the note as the first message. Ask the agent to repeat the task and the next step before it edits anything. In Cursor, a chat that already has messages is the wrong place. Start a new one.

Or choose Continue

BuildGist does not detect that limit. You open the other agent yourself. When you start a new chat in Claude Code or Cursor in the same project, a BuildGist card at the bottom right offers the earlier agent's task. Choose Continue and the new chat starts with that task's context. Choose Start fresh to skip it. Nothing is handed over unless you choose Continue.

The context is your own words about the task. The transcript stays in Codex. The diff stays in the folder, where the next agent reads the files. Claude Code takes the context on a new session, not a chat you are in the middle of. Cursor's handling of it is best effort: if the new chat does not show the task, paste the note. By default the card offers a task active within the last 12 hours. Codex needs you to approve BuildGist once, in Codex. BuildGist shows the steps. Setup is in Getting started.

When you choose Continue, the agent you continue in receives the task's context under your own account with it and its provider's terms. BuildGist records which agents that was between, and when.

Back in Codex after the reset

After the time in the message, you can return to Codex in this same folder. Tell it the task, what the other agent finished, and what is left. How a task moves among all three agents is in how to keep context across Claude Code, Cursor and Codex. If Claude Code is the one at its limit, see you’ve hit your session limit in Claude Code.

What you need

Codex needs you to approve BuildGist once, in Codex. BuildGist shows the steps. Setup is in Getting started.

To explain a run, your computer sends BuildGist what that run needs, and BuildGist passes it to Anthropic. BuildGist does not keep this material. It does not send your whole repository, files the run did not change, or files named like secrets, such as .env files and keys. It keeps the explanation and short cited excerpts of changed code, at most 12 lines each. Secret masking is best-effort. Privacy has the full list.

Every account gets 7 days of BuildGist Pro free, with no card. The 7 days start at your first explanation or Continue handoff. Nobody is charged when the trial ends. After that, BuildGist Pro is $12 per month, plus any applicable tax shown at checkout. See pricing.

Try BuildGist

Apple silicon, macOS 12 or later. United States, 18 or older.

Related

Questions: kayvanandre@gmail.com.