Application: Easily Switching Between Cloud and Local Models
These use cases make the Cost Control for Running AI Agents and Worry-Free, Local Operation of AI Agents solutions concrete. The scenarios are illustrative and show how DAOS works day to day.
At a glance
The problem
Cloud models and local models today mostly run in separate tools with their own login and their own rules. Anyone who wants to switch depending on the task ends up switching interfaces too – or sticks with a single option out of convenience, even when it isn't the cheapest or safest one.
The customer
Employee in day-to-day work
Uses a powerful cloud model for some tasks and a local model on their own machine for others.
The system
Open WebUI with the DAOS Policy Engine
One interface for every approved model, whether in the cloud or on your own hardware.
The goal
Use the right model for each task without switching tools
The speed and capability of the cloud where it fits – the control and cost of your own hardware where that's enough.
The result
In Open WebUI, every approved model sits side by side in one dropdown – no separate tools, no extra sign-in.
The path at a glance
One interface
Model chosen
Clearance checked
Processed where it fits
Answer with origin
Illustrative path – the detailed walkthrough is below.
The journey from the customer's side
- 1
One interface for everything
The employee opens Open WebUI like a familiar chat – cloud models and local models are both available in the same window, not in separate applications.
- 2
Pick the model the task needs
For research, they pick a powerful cloud model; for work with internal documents, a local model on their own machine or the company's internal server – one click in the dropdown is enough.
- 3
Clearance runs along automatically
In the background, the Policy Engine checks whether the chosen combination of model and data is allowed. Cloud providers that aren't cleared for critical data don't even show up as an option.
- 4
Processed where it belongs
If the request runs locally, no information leaves the employee's own machine or server. If it runs in the cloud, only what's approved for that goes out.
- 5
An answer with its origin
The answer appears together with which model produced it – the employee always knows whether they're working locally or in the cloud, without having to ask.
Further scenarios
Using hardware you already own
When the local model runs on machines the company already has, or on an internal server, these requests create no additional inference cost.
Configured centrally, used locally
Which cloud providers and local models show up in Open WebUI at all is something IT sets once, centrally, through the Policy Engine – per machine or per user group.