kayetana or cloud AI?
With cloud AI, every prompt goes to the provider. With kayetana, it stays on your network. Both have advantages, and here they are side by side.

Your own server versus the cloud.
| kayetana | Cloud AI services | |
|---|---|---|
| Where models run | Mac Studio on your network | Provider's data center |
| Where prompts go | They stay in-house | Over the internet to the provider |
| Prompts stored | Never | Depends on provider and contract |
| Works offline | ✓ | – |
| Very large models | Limited by memory | ✓ |
| Own hardware needed | Yes, a Mac Studio | No |
| Billing | Flat fee for the platform | Often per token or per user |
| Model choice | Open models, add your own | Provider's selection |
| Maintenance | Update script, nightly backups | Handled by the provider |
Cloud AI services vary a lot. This column describes the typical model, not any specific provider.
Your data never leaves the building.
kayetana processes prompts, documents and recordings on your own machine. Prompts and answers are never stored. It only logs time, endpoint, model, status, tokens and duration.
- Set your own retention periods for logs and datasets
- Automatic deletion when they expire
- No internet needed once the models are loaded
What speaks for each side.
kayetana: control
Models, documents and logs live on your server. That helps you with data protection.
kayetana: fixed costs
One flat fee for the platform, no matter how many tokens your team uses.
kayetana: offline
Chat, search and transcription keep running without an internet connection.
Cloud: large models
Models that need more memory than a Mac Studio has are easier to find in the cloud.
Cloud: no hardware
Nothing to set up, and no hardware, backups or updates to look after.
Cloud: light use
If you only use AI occasionally, pay-as-you-go billing can cost less.
See the difference for yourself.
In 30 minutes we'll show you kayetana with your own documents.
Frequently asked questions
Is kayetana as good as the big cloud models?
It depends on the task. For questions about documents, classification and transcription, the bundled open models often do the job well. For very large models, the cloud has the edge.
Are my prompts stored anywhere?
No. kayetana never stores prompts or answers. It only logs time, endpoint, model, status, tokens and duration.
Does kayetana make me GDPR compliant?
kayetana supports you with data protection because your data stays in the building. That is not a guarantee or legal advice, and the assessment is up to you.
Can I switch over from a cloud service?
Yes. kayetana offers a REST API under /v2 with chat, embed, rerank and more. The Playground shows sample code in Python, TypeScript and curl for every request.