Skip to content
onagero
DEBook a demo
Menu
Book a demo

Onagero

How Onagero worksHow a start worksPricingData securityContact

Solutions

ProductionControllingSalesPurchasingQualityOrder processing

Industries

MetalworkingAluminium and ProfilesMechanical EngineeringPlastics ProcessingFoundryAutomotive Suppliers

Integration

SAPproALPHAabas ERPMicrosoft DynamicsOracleExcel lists

Comparison

Onagero or ChatGPTOnagero or PalantirOnagero or Copilot StudioOnagero or RetoolOnagero or n8n

Knowledge

What does AI cost in mid-sized industry? All cost items, calculated honestlyAI and GDPREU AI Act in mid-sized industryIntroducing AI with the works council, not against itData masking for AIOn-premise AI or cloudGlossaryBook a demo

On-premise AI or cloud: what really stays in-house

Onagero, editorial team · published on · 6 min read

In short

“On-premise AI” can mean two things: the platform runs in your own premises, or the language model runs there too. With Onagero, the platform with your data, tools and permissions runs on your server; the language model runs at the model provider. If you need both in-house, you should clarify this before choosing any provider.

Two things that are often mixed up

An AI application consists of two parts: the platform, which reads data, provides tools, manages permissions and logs, and the language model, which understands and generates text.

Running the platform in-house is quite feasible. Running a high-performance language model in-house requires dedicated hardware, operations and compromises on quality. Many providers say “on-premise” and mean only the first.

What lives where with Onagero

  • On your server: the platform, the analysis copy of your data, the tools, roles and permissions, logs, backups.
  • In your systems: your original data. Onagero only reads.
  • At the model provider: the content of individual requests, in approved and masked form.

With Onagero, platform, data and tools run on your server; the language model receives only masked excerpts of individual requests. The complete data path.

When cloud is the better choice

  • You have no server infrastructure and do not want to run any.
  • Your data lives in cloud systems anyway.
  • You need many standard integrations, for example in Microsoft 365. See Copilot Studio compared.

When your own server is the better choice

  • Your data lives in ERP, shop-floor data collection (BDE) and databases on your own network.
  • Your IT policies require that data holdings do not leave the company.
  • You want to keep control over backup, recovery and access.

This article gives orientation and is not legal or tax advice.

Questions on this topic

Does Onagero need a graphics card?

No. Because the language model runs at the provider, a standard server or a virtual machine is enough. We state the exact requirements before the contract.

Does Onagero work without internet?

The tools calculate without a model and keep running without a connection. Functions that need the language model require a connection to the provider.

What happens if the model provider has an outage?

Tools that only calculate and display keep running. Questions to the model are then not possible. How we handle outages is set out in the operations contract.

Read on

More articles

Solutions

Which problem would
your team solve first?

Show us a real process. We will show you how it could look with Onagero.

Book a demo →

Check first whether a pilot fits: fit check in 2 minutes →