Guide

Claude Code and Codex in the cloud: how the limits work and where overflow work can go

Both run in the cloud already. The catch is that they share limits with the rest of your plan. Here is how that works and what to do about it.

By Mark Fulton, founder of AGI LayerUpdated Competitor facts checked
The short answer

Claude Code and Codex both offer cloud sessions that keep working after you close your laptop, and both draw on shared plan limits: Claude Code shares them with Claude chat on five-hour and weekly windows, and Codex shares one allowance with ChatGPT Work. When you hit a limit you can wait, upgrade, or buy extra usage at API rates. A separate workspace such as AutoAGI, which runs Claude Code, Codex, Hermes, OpenCode and DeepSeek in Cloud Harnesses on its own weekly dollar allowance, gives overflow work somewhere to go without touching those limits.

Read this first

AutoAGI is our product. The Claude Code and Codex facts below come from Anthropic's and OpenAI's own pages, checked on September 29, 2026 and linked under Sources. Most of those pages show no date, and limits change often, so confirm current numbers with the vendor.

How Claude Code works in the cloud

Anthropic describes Claude Code as an agentic coding tool that reads your codebase, edits files and runs commands, available in the desktop app, IDEs and the browser. Cloud sessions, started from the web, mobile or the desktop app, run on Anthropic's infrastructure. Per Anthropic's documentation, "the session keeps running after you close your laptop", and you can check on it or steer it from any device.

  • Plans. Claude Code is included on Pro ($20 a month), Max 5x ($100) and Max 20x ($200), and on Team and Enterprise plans.
  • Shared limits. Pro and Max plans share usage limits across Claude on the web, desktop and mobile and Claude Code. Cloud sessions share rate limits with the rest of your account, with no separate charge for the cloud VM.
  • Windows. Usage resets on a rolling five-hour window, and Max plans also have a weekly limit that applies across all models.
  • Idle expiry. Cloud sessions stop after a period of inactivity and the VM is reclaimed. Reopening restores history on a fresh VM, but background work still running at that point is not restored.

How Codex works in the cloud

Codex cloud tasks each run in their own sandbox, preloaded with your repository. OpenAI states that Codex and ChatGPT Work share one allowance and credit pool, that local messages and cloud chats share your plan's usage allowance, and that weekly limits may also apply. Codex is included across ChatGPT plans, including Free and Go with lighter limits, with Plus at $20 a month and Pro from $100 raising them.

During the agent phase, internet access is off by default unless the environment enables it.

The three side by side

Claude Code cloudCodex cloudAutoAGI Cloud Harness
Included onPro $20, Max 5x $100, Max 20x $200, Team, EnterpriseAcross ChatGPT plans, with higher limits on Plus $20 and Pro from $100Founder $29, Pro $100, Max $200
AllowanceShared with Claude chat, five-hour windows plus a weekly limit on MaxShared with ChatGPT Work, allowance windows plus weekly limitsIts own weekly allowance in dollars, $6.70, $23 or $46
Beyond the limitWait, upgrade, extra usage at API rates, or API creditsBuy additional credits, upgrade, or use an API keyMove up a membership, or usage credits at face value when on sale
EnginesClaude CodeCodexClaude Code, Codex, Hermes, OpenCode and DeepSeek
OutputChanges in your repositoryChanges in your repositoryFiles in your library and, with a GitHub connection, pushes to a repository
Beyond codeCoding focusCoding focus, plus Work in the same allowanceResearch, writing, analysis, Jev scoring and 51 connected apps

Anthropic and OpenAI figures from their own pages on September 29, 2026.

What to do when you hit a limit

  1. Wait for the reset. Five-hour windows reset quickly. Weekly limits take longer.
  2. Buy extra usage. Anthropic offers usage credits billed at standard API rates for Pro and Max. OpenAI sells additional credits on Plus and Pro.
  3. Upgrade a tier. Max and Pro tiers raise your limits, at a higher monthly price.
  4. Move overflow work to a separate workspace. Work that does not need your own local repository, such as a prototype, a research-backed build or a scheduled report, can run in an AutoAGI Cloud Harness on a different allowance.

Using AutoAGI as the overflow lane

Every AutoAGI membership includes every Cloud Harness. A harness run gets its own workspace and session, is metered on what the agent actually used within a compute ceiling that rises with your plan, and lands its output in your library. A run that hits its time limit keeps its workspace so the next one continues.

Connect GitHub and the run can push its files to a repository. Cloud Harness runs reach your apps through the AutoAGI gateway and never hold a provider token. Read how connected apps stay safe.

When not to use it. If you want to work interactively inside your own local repository with your own configuration and tools, use Claude Code or Codex directly. AutoAGI suits delegated jobs that can finish on their own.

Questions

Does Claude Code keep running when I close my laptop?
Cloud sessions do. Anthropic says the session keeps running after you close your laptop, and you can check on it from any device. Sessions stop after a period of inactivity and the VM is reclaimed, though history is restored when you reopen.
Do Claude Code and Claude chat share usage limits?
Yes. On Pro and Max plans, Anthropic states that usage limits are shared across Claude on the web, desktop and mobile and Claude Code, including cloud sessions.
Do Codex and ChatGPT Work share an allowance?
Yes. OpenAI states that ChatGPT Work and Codex share usage, and that local messages and cloud chats share your plan's usage allowance, with weekly limits that may also apply.
What can I do when I run out of Claude Code usage?
Wait for the reset, upgrade to a higher tier, enable extra usage billed at API rates, or use API credits. Some work can also move to a separate workspace such as an AutoAGI Cloud Harness on its own allowance.
Can AutoAGI run Claude Code and Codex?
Yes. Claude Code, Codex, Hermes, OpenCode and DeepSeek are available as Cloud Harnesses on every AutoAGI membership, on a weekly dollar allowance separate from your Claude and ChatGPT plans.
Should I cancel Claude Code or Codex if I use AutoAGI?
No. Keep them for interactive work in your own repository. AutoAGI adds delegated cloud execution on a separate allowance.

Sources

Keep reading

Related comparisons and guides

Compare

AutoAGI vs ChatGPT: where an autonomous workspace beats an assistant, and where it does not

AutoAGI vs ChatGPT Work. ChatGPT Plus is $20 and runs OpenAI models on a shared allowance with Codex; AutoAGI is $29 to $200 with Claude, OpenAI, Gemini, Grok and weekly usage in dollars.

Guide

Autonomous AI agents: what they are, how to choose one and what they cost

A plain guide to autonomous AI agents: a clear definition, how they differ from chat assistants, the criteria that matter when choosing one, how usage is metered and which products fit which job.

Pricing

AutoAGI pricing: Founder, Pro and Max

AutoAGI memberships start at $29 a month. Every plan includes every model, every Cloud Harness, Jev analysis and 51 connected apps. Usage is stated in dollars and resets every 7 days.