Guide
Auto-continue Codex CLI after a usage limit
How to get Codex CLI to carry on by itself when its usage limit resets, using a free macOS menu-bar app instead of a terminal wrapper.
Read the guide: Auto-continue Codex CLI after a usage limit
macOS menu-bar app · for coding agents
Your coding agent hits its usage limit at midnight. The limit resets at 3am. You are asleep. Agent Auto-Continue sends the “Continue” for you, right on time.
Free and open source. View it on GitHub
Conversations from T3 Code, Claude Code, Codex, OpenCode and the Claude and ChatGPT desktop apps, in one list, most recent first.
Usually just “Continue”. Anything you would type into the thread yourself works.
Quick picks for +5 min, +30 min, +1 hour or tomorrow morning, an exact date and time, or as soon as the usage limit resets.
Sending a message to an agent twice is worse than not sending it at all, so the app is careful.
Short pages on the things people actually search for, plus the docs.
Guide
How to get Codex CLI to carry on by itself when its usage limit resets, using a free macOS menu-bar app instead of a terminal wrapper.
Read the guide: Auto-continue Codex CLI after a usage limitGuide
Claude Code now carries on by itself after a usage limit resets: what the built-in setting does, what it skips, and when an app still helps.
Read the guide: Claude Code auto-continue: what's built in, and what it doesn't doGuide
Why your Mac sleeps mid-task, when caffeinate is all you need, why a closed lid still wins, and how to stay awake only while agent work waits or runs.
Read the guide: Keep your Mac awake while a coding agent runsDoc
Install Agent Auto-Continue on your Mac: open the disk image, answer macOS once, check your agents in Settings and queue your first Continue.
Read the doc: Install and first runDoc
The six coding agents Agent Auto-Continue can schedule, how it reaches each one, and which know when a usage limit resets or work with the Mac locked.
Read the doc: Supported agents and what each can doGuide
Your coding agent stopped mid-task: when to type Continue in the same thread, when to start fresh, and a handoff prompt to carry the work across.
Read the guide: Continue the thread or start a new one?Guide
When plain Continue is enough for a coding agent, when a specific message works better, and how a stop phrase lets a run of turns end itself.
Read the guide: What to say instead of ContinueGuide
How to make a Codex thread in the ChatGPT desktop app carry on by itself when its usage limit resets, and what that asks of your Mac overnight.
Read the guide: Auto-continue Codex in the ChatGPT appGuide
Claude Desktop can retry a turn once the session limit resets; here's what that checkbox covers, and what a free menu-bar app adds on top.
Read the guide: Auto-continue Claude Desktop sessionsGuide
How to leave Claude Code or Codex working overnight and wake up to progress: a turn limit, a sane permission mode, no double-sends, a morning review.
Read the guide: Running an agent overnight without breaking thingsDoc
Queue a Continue from your phone over Tailscale, or let Claude Code or Codex schedule their own, with Agent Auto-Continue's opt-in remote control.
Read the doc: Control it from your phone or another agentDoc
Why Agent Auto-Continue didn't send your Continue, what an unconfirmed delivery means, why a continuation paused, and how to grant Accessibility.
Read the doc: FAQ and troubleshootingGuide
Where Claude Code shows when your usage limit resets, how the five-hour and weekly windows work, and how to pick the task back up right at the reset.
Read the guide: When the Claude limit resets, and how to pick up right thenSupport this app
Agent Auto-Continue is a side project. If it saves you a night of idle agent time, a coffee keeps it moving.
Support on Ko-fiOpens ko-fi.com/velvetacorn in a new tab.
Logo moon: Albrecht Altdorfer, The Battle of Alexander at Issus (detail of the crescent), 1529. Alte Pinakothek, Munich, inv. 688. Public domain scan, Google Art Project via Wikimedia Commons.