Nanoodle: static AI apps powered by your users' NanoGPT balance
Drop a "Sign in with NanoGPT" button into any static HTML page and let users spend their own NanoGPT balance on your workflow with zero backend.
Nanoodle makes that concrete. You build a workflow in the browser, run it with NanoGPT, then turn it into a shareable app. The exported app is just HTML and JavaScript, so it can be hosted anywhere a static page can live.
The surprising part is not that Nanoodle uses OAuth. It is that the exported app does not need its own inference backend, key storage, provider account, or billing integration. The user signs in with NanoGPT, approves access, and the app runs against that user's NanoGPT balance.
A workflow that becomes an app
Imagine a research brief generator with four nodes:
- A chat model turns a topic into a structured research brief.
- An image model turns the brief into a clean diagram.
- A video model turns that diagram into a short presentation clip.
- An audio model reads the final summary aloud.
In a normal AI app, that means a server, API keys, model routing, usage tracking, and some way to charge the person running it. In Nanoodle, the chain runs in the browser. Click Create App and it becomes a single static HTML file. Share that file or host it on a static site, and the next person can sign in with NanoGPT and run the same workflow with their own balance.
That is the shift: a workflow is not trapped inside the builder, and the exported app is not a toy demo that stops working the moment it needs paid inference. It can call NanoGPT because the user brings the account, the balance, and the approval.
What changes
Before, shipping even a tiny AI tool meant deciding where secrets live, how requests are metered, and who pays the model bill.
With Sign in with NanoGPT, the static app can request a dedicated NanoGPT key for the signed-in user. NanoGPT handles the models, routing, usage, and balance. The user keeps control from NanoGPT, including spending limits and revocation.
Try it
Open nanoodle.com, sign in with NanoGPT, build a workflow, and use Create App to export it as a static app.
For implementation details, see our Sign in with NanoGPT OAuth guide.
