Don't let one AI company set your terms.
Prices rise, limits shrink, models vanish and privacy policies change. Your Own Stack keeps every major model one tap away on your own keys, so you can switch the day the deal gets worse.
When you rely on one AI app, the deal is theirs to change.
Plans get repriced and usage limits get cut, sometimes before anyone announces it. You find out mid-task.
The model you've tuned your prompts and habits around can be retired from the app on a couple of weeks' notice.
A chat app can start training on your conversations by default and keep them for years, unless you notice and opt out.
Months of chats and context live inside one company's app. Leaving means starting over.
- PrivacyAug 2025: a major chat app began training on consumer chats by default, keeping them up to five years for anyone who didn't opt out. Source
- ModelsFeb 2026: a popular model was pulled from a leading chat app about two weeks after the announcement. Source
- LimitsA major provider cut subscribers' peak-hour session limits and confirmed it only after users noticed. Source
When you're locked into one app, every change is theirs to make and yours to absorb. You can't take your chats with you, trying a rival means starting from scratch, and saying no isn't really an option. Unless switching costs you nothing.
Keep every model within reach, on your own keys.
Switch in one tap
Every model sits in the same chat. If one gets pricier, move to another without a migration. You pay providers per use and see each message's cost first.
No single point of failure
Your workflows aren't welded to one model. Swap the model behind any step and keep working.
API terms, not app terms
You connect through each provider's API with your own key. Paid API plans generally don't train on your messages by default.
Your history stays put
Chats are saved in your browser, not in our cloud, and stay with you whichever model you use next.
Your stack on every device, through storage you already trust.
Back up your keys and the chats that matter to your own cloud storage. We never hold a copy.
Encrypted before it leaves
Your keys are locked with a password you choose before they're uploaded. Without it, the file is unreadable, even to your storage provider.
You pick what syncs
Sync your keys, the chats you mark as important, or nothing at all. Everything else stays on the device.
Your storage, not ours
The backup lives in your own account. Switch storage providers whenever you like, the same way you switch models.
When the next price hike lands, you're one menu away.
Start with Writing, Coding or Research, or build your own. Assign a model to each step.
Connect only the providers you use. A test button confirms each key works.
Use the best model for each step today. If one changes for the worse tomorrow, swap it and carry on.
One price, once. We can't change the deal on you either.
- 100 messages total
- 1 workflow
- Chats saved on this device
- Unlimited messages and workflows
- Encrypted sync to your own storage
- No subscription, ever
Before you start
Do my messages pass through your servers?
No. Messages go straight from your browser to the provider you pick, under that provider's API terms.
Where are my keys stored?
In your browser. Save them to your own cloud storage and they're encrypted with your password first. Either way, they never reach us.
Do I need API keys?
Yes, one per provider you want to use. The app links to each provider's key page and tests the key for you.
What will the models cost me?
You pay each provider per use. The app shows every message's cost before you send it.
Which models can I use?
Claude, GPT, Gemini, DeepSeek and Mistral directly, plus many more through OpenRouter.
Can I run local models?
Not yet. A desktop app that runs local models is planned.
Keep the power to walk away.
Set up your stack once. Use the best model today, and switch the moment it stops being the best deal.