About omesos
Private AI in your cloud, not ours.
omesos sets up private AI inside your own AWS cloud: we choose the models that fit your use case, run them under your keys in your region, and look after them from then on.
01 · What we do
Set up by us, run by us, owned by you
We work for software vendors who want AI features in their product, on models they control, and for companies that want a private assistant for their internal documents.
There are hundreds of open-weight models, and new ones appear every month. Choosing well is the core of our work: we test candidates on your own examples, in German and English, with your budget in mind, and run the best of them in your cloud, on Amazon Bedrock, the managed AWS model service, or on dedicated GPU instances, whichever fits.
Setting it up is only the start. We also operate the system, keep it current and upgrade the models as better ones appear.
02 · What we believe
Principles we build on
They shape every system we set up and every recommendation we make.
Who can see what, in detailYour data stays in your cloud.
The system runs in your own cloud, in the EU region you choose, for example Frankfurt. You own the keys, and you can revoke our access at any time.
No AI vendor in the data path.
Your prompts and documents are not sent to any AI model vendor. We bring the model to your data, not your data to a model vendor.
The right model, not one model.
No single model is best at everything. We choose by use case, language and budget, for example from the Llama, Mistral, Qwen or Gemma families, test the candidates on your own data, and move to something better when it appears, once you have approved the change.
Everything in writing.
Before we work in your cloud, we sign a GDPR processing agreement under Art. 28 GDPR and a separate secrecy agreement with you. Our access is time-limited and logged in your own CloudTrail. We also help with the paperwork around it: the AI notice for your users, a licence register for the models and input for your DPIA.
03 · How we work
From first call to daily operation
Assessment
Thirty minutes on your use case, your data and your AWS setup. You leave with a concrete recommendation for the next step.
Proof of concept on your data
We shortlist candidate models, build a test set from your real examples and run the candidates in your cloud. You see how each one handles your tasks, in German and English, before you decide anything.
Production
The chosen setup goes live in your own cloud: a private endpoint for your product, or a chat interface with search over your documents, rolled out to the people it is meant for.
Operations, with a named contact
We monitor the system, keep it up to date and bring in better models after you approve them. You have a named person who knows your setup, with support during business hours (CET) and a reply within 2 working days.
04 · Independent
Paid only by our customers
Nobody pays us to recommend a model, a cloud service or a tool. Every recommendation is about your use case.
Not a reseller. Your AWS bill comes directly from AWS. We do not resell cloud capacity or add a margin to your usage.
No commissions. We take no commissions, referral fees or other incentives from AWS or from any model vendor.
No favourite model or setup. Which models you use, and whether they run on Amazon Bedrock or on dedicated GPU instances, follows from your use case, your data and your budget.
Start with your use case
A 30-minute assessment, no slides. Bring your use case and your AWS setup; you leave with a concrete recommendation for the next step.
Book an assessment