lupy.io
Sections
EN
Download

Using your Claude and Codex accounts as one team

If you pay for more than one AI account, you can make them work together instead of switching tabs. A manager, a crew and a second opinion fit together.

Published: 7 min read

Plenty of people now pay for more than one AI account. Maybe two Claude accounts, one for work and one personal, plus a Codex account. Each has its own limits, its own strengths, and its own tab. In practice you pick one, run into a limit, and move to the next by hand.

This post describes a more orderly way to use them: treat the accounts as members of one team, with different roles, working on the same project.

The problem with one account at a time

Working with one account feels simple until a job gets long. Two things happen. First, an account can run out of room for the moment and make you wait. Second, one model is used for everything, including small chores that do not need the strongest option.

Switching by hand fixes neither problem well. You lose the thread of what was done, you re-explain the project, and you start to keep notes about which account knows what.

A manager and a crew

The pattern that works better is the one human teams use. One person plans and coordinates. Others do the focused work. In lupy.io, you put the strongest Claude model on the manager, the part that reads your request, plans and hands out tickets. The specialist agents are then spread across your other accounts, your Claude accounts and your Codex account.

The spread of agents over accounts is something you set up once, not something you renegotiate every time you open a new chat. That is the real gain: the arrangement lives in the project, so you stop carrying it around in your head.

A human comparison makes it clearer. A small agency does not ask its most senior person to proofread every invoice. The senior person decides the approach, and the rest of the crew handles the volume. The same logic applies here, with the strongest model reserved for the planning that shapes everything downstream.

Accounts that cover for each other

Accounts rest. A limit is reached, a session pauses. When that happens in lupy.io, the job carries on in the next account instead of stopping. In the Accounts window you can see which one is working, which one is resting and which one is ready. When a job moves, a short notice tells you it moved to the next account.

This is not magic and it does not make limits disappear. It means a pause on one account does not have to become a pause on the whole project.

A second opinion on tap

There is a third use for several accounts that people overlook: checking each other. You can have another service double-check an answer. If one agent drafts a report, a different one can read it with fresh eyes. The reason this helps is mundane. Two readers catch different mistakes.

No credits, no tokens

lupy.io runs on the accounts you already own. It does not sell you credits, and there is no token meter on lupy.io's side. What you pay lupy.io for is the lupy.io plan: a monthly or yearly subscription, or a one-time lifetime license. The usage of your Claude and Codex accounts stays on those accounts, under the terms you already have with each provider. See the pricing page for the current plans.

A walk-through

lupy.io is a Mac app, and what you work with are windows, task cards and an inbox. The same screens show you the plan, the tickets and the state of each account, so you are never guessing which part of the team is busy.

A small tip for the first run: start with a job that has a few independent parts, such as research plus a draft plus a summary. Independent parts are where a team shows its value, because they can move at the same time.

A realistic scenario

Say you have two Claude accounts and one Codex account. You give the manager the strongest Claude model. You spread research and writing jobs to one account and code jobs to the other. When one account goes to rest in the afternoon, the work continues on the next one, and the Accounts window shows you that.

You do not have to keep three separate tabs open, remember where each one stopped, or carry the context by hand. These are the ordinary, everyday benefits of having the arrangement written down inside the project.

Why it is worth the small setup

Connecting accounts takes a few minutes, and it pays back the first time a long job runs past a pause on one account. Without a shared setup you notice the pause, switch tabs, paste context and hope nothing was lost. With one, the project simply continues on the next account.

There is also a calmer effect that is easy to underrate. When the arrangement is written down in the project, you stop keeping a mental map of which account is good for what.

The most common mistake

The most common mistake is to pile all the agents onto one account. That does not solve the problem; it only makes a single tab look like many. The point of spreading is to spread the load and the risk of the work.

The second mistake is never using the second opinion. Having another service read what the working agent wrote is often a few minutes of effort, and it can catch an error that would otherwise go unnoticed.

Who this suits

It suits people who already pay for more than one account and feel the friction of juggling them. It also suits people with one account who expect to add another later; the setup does not change, you just have more seats on the team.

It does not suit anyone who wants a single chat with a single model and nothing more. For that, the chat window of the provider is simpler, and we say so plainly in A chatbot or a team of AI agents?.

What stays on your side

Your project files stay on your Mac. lupy.io does not keep them in its own cloud, and they are not deleted when a membership ends. Connecting accounts does not move your work somewhere else.

If you want the full list of how it behaves, the FAQ covers accounts, files and licenses, and the next post in this series, How to write a request AI agents can finish, shows what to type once the accounts are connected.

In short

Running several accounts as one team has three concrete benefits: work does not stop when one account pauses, the strongest model is kept for planning, and you can have another service check what an agent wrote. All of them rest on the same idea: set the arrangement up once, then stop thinking about it again and again. In return, the only thing you pay lupy.io for is the plan; the usage of your accounts stays with their own providers, under the terms you already accepted. Looking at the prices and the FAQ before you start helps you set the right expectations. Knowing in advance what happens when an account reaches its limit also keeps you calm when you begin a long job. Once the arrangement is in place, you can give your attention to the work itself rather than to the accounts.

If you want the full list of how it behaves, the FAQ covers accounts, files and licenses. The next post in this series, How to write a request AI agents can finish, shows what to write once the accounts are connected.