How to use your Claude or ChatGPT plan with other apps

Subscription or API key: what’s the difference?

There are two ways to pay for Claude or OpenAI’s models. A subscription (Claude Pro, ChatGPT Plus and so on) is a flat monthly price with usage limits, tied to your personal or work account. An API key is pay-as-you-go billing per token, meant for developers building software. Many AI apps ask for an API key, which means a second bill on top of the plan you already have. Using your subscription avoids that.

Which plans include Claude Code and Codex?

As of the date on this page, from Anthropic’s and OpenAI’s help pages. Features and usage limits differ by plan; check the current pages before you buy.

Do they share limits with the chat app?

For Claude, yes: Anthropic says the limits are shared across Claude and Claude Code, so all activity in both counts against the same allowance. For Codex, OpenAI describes it as included in your ChatGPT plan with limits that vary by plan. In practice, heavy agent use can use up an allowance faster than chatting, because an agent makes many model calls per task. Higher tiers (Claude Max, ChatGPT Pro, Team premium seats) exist largely for that.

How can another app use your plan?

The straightforward way is for the app to run the official agent, Claude Code or Codex, on your computer, where you have signed in with your own account. The agent talks to Anthropic or OpenAI as it would if you used it in a terminal; the other app gives it a task, shows its progress and adds things around it. Your sign-in stays with the official tool.

Be wary of anything that asks for your Claude or ChatGPT password, or copies your login to its own servers. Read each provider’s terms for what they allow, and prefer apps that run the official tools locally.

Step by step: using your plan

1. Check that your plan includes the agent (table above).
2. Install the official agent: Claude Code from Anthropic or Codex from OpenAI.
3. Sign in with your Claude or ChatGPT account when it asks, choosing the subscription sign-in rather than an API key.
4. Use it directly, or in an app built on it, which will find your signed-in agent.
5. Watch your usage in your account settings; if you hit limits often, a higher tier or a lighter model may help.

Limits and gotchas

• Usage caps. Subscriptions have them; long agent tasks reach them sooner.
• Work accounts. On Team and Enterprise plans an admin may control who gets the agent.
• Models change. Which models a plan offers changes over time, and older ones retire.
• Automation. API keys suit unattended servers and shared environments; OpenAI points to API keys for things like CI. A subscription is for you, on your machine.

Where AppuBot fits

AppuBot is free during early access and needs no API keys: it runs Claude Code, Codex, Cursor, Grok, GitHub Copilot or OpenCode with the account you’re already signed in to on your computer. You can add more than one account of the same agent, like a work and a personal Claude, and choose which bot uses which. Its Usage page shows how much your agents were used, read on your computer. More in your AI account.

Sources

Facts about other products are from their makers’ pages as of 29 October 2026 and may have changed since. Product names are trademarks of their owners.

• Claude Help Center: Using Claude Code with your Pro or Max plan
• Claude Help Center: Using Claude Code with your Team or Enterprise plan
• OpenAI: Codex pricing and plans