You can manage AI subscriptions in one app with Aiola, running each official CLI on the plan you pay for. Each thread picks its provider and model, while Usage shows the limits and reset times each provider reports. Aiola adds no token markup; your subscriptions stay separate, under each provider's own terms and limits.
What does managing several subscriptions actually mean?
If you pay for several coding plans, the daily problem is deciding which agent should take the next task. You need the right account signed in, a model selected, and enough reported quota to begin. Repeating project instructions in separate tools adds another step.
Aiola brings those choices into one desktop app. It supports Claude Code, Codex, Cursor, Grok CLI, OpenCode, and Antigravity on your own subscriptions. The supported providers page lists the tools you can bring into the workspace.
The subscriptions remain yours. Aiola does not pay for, pool, or resell them, and it does not resell models. A Claude allowance does not become a Codex allowance because both appear in the same app. Each provider's own terms and limits still apply.
Think of management here as setup, provider selection, and usage visibility. Keep purchases, renewals, and account questions with the relevant provider. Check each provider for current plan prices, included CLI access, and limits before deciding what to buy or keep.
You can use Claude and ChatGPT subscriptions together through their separately authenticated CLIs. That means Claude Code uses its account, while Codex uses its OpenAI or ChatGPT account. Signing into Aiola does not sign you into either CLI.
How do you connect the plans you already have?
Start with the accounts you intend to use for coding. Install each provider's CLI on the desktop where Aiola runs, then authenticate in a terminal. Cursor is the exception: you sign in with Cursor inside Aiola. The Providers page does not accept provider API keys.
For the four providers in this guide, sign in like this:
| Provider | How to sign in |
|---|---|
| Claude Code | claude login |
| Codex | codex login |
| Cursor | Sign in with Cursor in Central Settings, Providers |
| Grok CLI | grok agent --reauth |
Open Central Settings, then Connections, then Providers. Select Re-check so Aiola can read each CLI's local authentication state. An installed provider must be authenticated before you can enable it.
Check the account details shown before starting work. The page can report the account email, subscription, installation state, version, binary path, and discovered models. A failed detection check is reported separately from an unauthenticated account.
Enable the providers you want, then expand each one to choose visible models and their order. Hidden models stay available on the settings page, but disappear from model pickers. Provider availability applies across the app.
Setup is local to the machine. Your coding agents run on your own desktop, rather than in a vendor cloud. Keep that desktop available when you intend to run them.
How do you choose the next provider when one runs out?
Open Central Settings, then Billing, then Usage before assigning the next substantial task. The Usage page brings provider reports together, including limits and reset times when available. Select the circular-arrow button beside the Usage heading to request current data.
Read each card as a report from that provider. Available fields differ by provider and account, so missing data does not establish that an account has unlimited capacity. Usage does not create one shared allowance across your plans.
Suppose Claude reports that its allowance is exhausted while Codex still has room. Open a new thread in the same workspace, and choose Codex and an available model before sending. Mention the same task or report so the next agent can work from the project evidence.
This is a deliberate choice for the next thread. Do not assume that a usage card transfers an existing conversation or arranges a handoff. Select the provider yourself, and give the new thread the request it needs.
A workspace default can preselect a provider and model for new threads. That default selects from globally enabled, visible models. Changing it does not install, authenticate, or enable another provider.
For card sources and refresh details, read tracking Claude Code and Codex usage limits. For two simultaneous coding threads, use the Codex and Claude Code side-by-side guide.
What stays with the project when you choose another agent?
A workspace is one project in Aiola. Its instructions, tasks, notes, and records stay available whichever provider you choose. A new thread can mention the same task, log group, or feedback item.
The multiple-provider workflow starts from that shared project context. Write the request once as a task, then use it when choosing another agent. State the expected result and the evidence the agent should inspect.
For example, a user report might describe a failed checkout. Open the relevant feedback and production error, then ask the selected agent to inspect that evidence. If you later choose another provider, point its new thread to those same records.
Aiola is one app to build, run, and support your products. App Logs, feedback, analytics, GitHub issues and pull requests, and tasks give agents the material they need. Agents can only fix what they can read. Keeping those records beside the code reduces copying context into another chat.
Hosted Analytics, App Logs, and Feedback are part of Pro. Provider subscriptions remain separate from that access. Choose the Aiola access that fits your workflow, and check provider plans with each provider.
Try it in Aiola
Give your coding agent a task and the project evidence it needs, then choose its provider and model. Download Aiola to use your existing CLI subscriptions and check reported usage before the next run. Start with one project and the providers you already pay for.
