About idapt

I built idapt because I was tired of three things, and I kept hearing the same three from people who work the way I do.

Tired of subscriptions that ration the work. On a Max plan, the weekly reset turned into a schedule: work until 3 a.m. to spend the allowance before it vanished, or stop mid-task because the counter hit zero. The work did not care that it was Tuesday.

Tired of the model treadmill. Every month a new model leads, and every switch meant a new subscription, a new tab, and the same context pasted in again. The best model for the job kept changing; my tools kept assuming it would not.

Tired of the fine print on my data. I write things for clients. I did not want my code and my conversations to be training material by default.

So I built the workspace I wanted:

• Every model, one place. Two hundred plus, one login. When a new model leads, it is a dropdown away, not another bill.
• Real work, with receipts. Agents run on real computers, files stay across chats, and the run log shows what happened.
• No rationing. Platform usage bills transparent credits at the prices shown in the app. Your own keys and local models are free lanes, always uncapped. A heavy week costs what it costs; it does not stop when a counter resets.
• Your work is not training data by default. On your keys or a local model, it runs entirely on your terms.

One thing said plainly: idapt is hosted, not end-to-end encrypted. If you want the strongest isolation, pair a machine you own and run a local model; the whole product works that way too.

Start free. No card, no trial clock. Bring your own keys, or add credits when you want the frontier models. If a paid plan is not for you, the free tier is permanent.

If any of this describes the work you actually do, I built this for you.