Bring your own AI to every app.
mycorra is one AI account you take with you. When an app shows Log in with mycorra, its AI features run on your account instead of yet another subscription. You get one balance, a spending limit for every app, and no technical know-how needed.
How it works
Three steps, no manual.
- 1
Make your account
All you need is your email, and there's no password to remember. Add credit once and it works in every app you connect.
- 2
Find the button
Apps that work with mycorra show a Log in with mycorra button. Tap it, check what the app asks for, say yes.
- 3
Just use it
Write, summarise, translate, brainstorm. The app's AI runs on your mycorra balance, within the limit you set.
What changes
One AI account instead of an AI subscription per app.
- Today
- Every app with AI features sells its own plan, whether you use it once a month or every day. Your AI spending ends up spread across a pile of separate bills.
- With mycorra
- Apps use your account. You pay for what you actually use, from one balance, and see every app's usage in one place.
One balance, every app
No separate AI subscription for each tool. Every connected app draws from the same account, and you only pay for what you use.
You're in charge
See every connected app, give each one a spending limit, and disconnect any of them with a click.
No tech skills needed
No API keys and no setup. Each app recommends a model, so you only pick one if you want to.
Private by default
Built in Germany under EU privacy law. Apps get AI answers, never your other apps or payment details.
For builders
Usage-based AI, carried by your users.
A flat AI subscription makes you guess. Price it too low and your heaviest users eat your margin; price it too high and the light users leave. With mycorra, AI is paid per use by the person who uses it. Add Log in with mycorra and every request runs on that user's own balance.
- Each user pays for exactly the AI they use
- No token budget to pre-buy, no margin risk, no metering to build
- Standard OAuth 2.0 sign-in and one documented inference endpoint
- A catalog of models you can rely on, and your users choose among the ones you recommend
// after "Log in with mycorra"
POST /api/inference
Authorization: Bearer <user's token>
{
"messages": [
{ "role": "user",
"content": "Plan three dinners with lentils" }
]
}
// metered and charged to this user's
// balance, not to your bill
Models: you recommend, your users decide.
- 1
Build on a catalog you can rely on
mycorra publishes a list of available models, each with a stable ID and a per-token price. You develop and test against that list instead of tracking a dozen providers.
- 2
List what works best for your app
Mark the models you've tested: the ones you recommend, and the ones that work but trade some quality for speed or price. That list is what people see when they connect your app.
- 3
Users pick the model
Each user chooses from your list, or simply keeps your top recommendation. Your code doesn't change: same endpoint, same request, whichever model they picked.
Recipe planner works best with
- Llama 3.3 70BBalanced, the app's pickRecommended€€
- Qwen2.5 72BStrong at structured listsRecommended€€
- Llama 3.1 8BFastest, lowest costWorks€
- DeepSeek V3Most capableWorks€€€
You can change this anytime in your mycorra account.
Questions
Good to know.
What does it cost?
You pay for the AI you actually use, from one prepaid balance, with no subscription. You give every app a spending limit, and nothing is ever charged beyond what you've topped up.
Why not just pay each app for its AI?
You can, and many apps will keep offering that. But it means a separate AI subscription for every tool, each with its own bill. With mycorra you have one balance, one place to see what each app uses, and one click to cut any of them off.
Which apps work with mycorra?
We're bringing on the first apps right now. People on the waitlist hear about them first, and builders on the list get early access to integrate.
Do I need to know anything about AI?
No. There's nothing to set up, and every app comes with a recommended model already selected.
Can I choose which AI model an app uses?
Yes. Each app lists the models that work best with it, and you pick one when you connect. Faster models cost less, more capable ones cost more, and you can switch anytime.
What happens to my data?
mycorra is made by Exaptation GmbH in Germany. We pass your request to the AI model and keep only what's needed to run your account. The details are in our privacy policy.
Is everything on this page available today?
Not yet. mycorra is in early access. Signing in with mycorra, paying from one balance, per-app spending limits and disconnecting apps already work. Letting you choose between the models an app recommends is still being built. Waitlist members get access as these parts open up.
When does it launch?
Soon. Join the waitlist and we'll send you a single email when your spot opens.