Use what you already pay for
If you already have a ChatGPT Plus subscription, a Claude subscription or an API key with any supported provider, you can connect it to OperatorNest and run your tasks on it. You pay the provider once for that model access, with no second markup from OperatorNest on the same calls. Compare the included tools and limits in ChatGPT Plus, Claude Pro and Google AI Pro before deciding which plans to keep.
This sits alongside model choice: you pick which model runs a task, and your key decides whose account pays for that model’s usage.
How connecting a key works
In your operator’s settings, add a key or connect a subscription for a provider, for example your existing OpenAI API key or your Claude account. From then on, any task you route to that provider uses your key. Connect keys for as many providers as you use and leave the rest on OperatorNest’s own access.
To remove a key, take it out of settings and it stops being used right away. Tasks already using it switch to an available alternative or pause until you choose one.
Where the key is used, and where it isn’t
A key you bring is used only for the model calls your own tasks make. It isn’t shared with other operators, pooled with other customers’ usage, or used outside the tasks you’ve routed to it. It’s encrypted at rest and in transit, and only your operator’s task execution touches it. You can see and remove every key you’ve connected at any time; see security for how OperatorNest handles access controls.
Seeing what your key costs
Usage from a key you bring shows up per task. For example, a research task that made a dozen calls to a reasoning model shows the model, the number of calls, and what that usage cost under your provider’s own pricing. This is part of every task’s receipt, so a week of recurring briefs is easy to add up. The AI subscription cost calculator compares plan totals and listed capabilities across subscriptions.
For outreach that needs a familiar writing model, see how an AI agent builds and checks a press list before drafting pitches for review.
If you mix your own key for one provider with OperatorNest’s access for another, the receipt shows which task ran on which account and what it cost.
Setting up a key correctly
Generate an API key or confirm a subscription is active in the provider’s own account settings. Paste the key into your operator’s settings, OperatorNest checks that it works, and it’s ready to use. If a key expires, hits a quota or is revoked on the provider’s side, your operator flags it on the next task that tries to use it.
What bringing your own key changes
Plenty of AI products charge you for model access you already pay a provider for directly. That’s a reasonable business model, and it isn’t OperatorNest’s default. If you’re a heavy user of one or two providers, your operator’s model usage rides on plans and pricing you’ve already negotiated for yourself, with no second markup on top.
You also hold the key. You can see the usage and revoke it the moment you want to.
A concrete example
Say you pay for a Claude subscription and have an API key for a second provider you use occasionally. You connect both. Your inbox operator runs on Claude and uses your subscription for every task. Your research operator is set to Auto, so it draws on the second key when Auto routes work there and on OperatorNest’s own access otherwise. At the end of the week, your receipts show which tasks used which key and what each one cost, with no bundled charge to reconcile.
Mixing your own keys with OperatorNest’s
You don’t need a key for every provider. A common setup is your own key for the model you rely on most and OperatorNest’s access for the rest, for example to try a model you don’t have a subscription for. Model choice, task history and memory work the same either way; only the account paying for that model call changes.