# univedo

> univedo is an integration platform for data and processes. It connects the systems, devices and people of an organisation into one shared, versioned picture. Applications from different vendors work on the same data and hand processes to each other. New applications, including ones generated from a plain-language prompt, reach production data only after an expert has approved them, and the platform enforces that approval at runtime. univedo runs in the cloud, on premise or in a private cloud, and an installation can switch between them.

Key facts:

- Category: integration platform for data and processes; also a runtime for approved, AI-generated ("prompted") applications.
- Deployment: cloud, on premise, private cloud. All three run the same software; data, applications and approvals move together when switching.
- Safety model: applications build on a shared data model, are developed in a separate space, every change is versioned, an expert approves them before they may write production data, and the platform checks that approval on every write.
- AI control: all AI calls from applications go through one configurable component; providers, models and limits are chosen by the operator, and models can run on site.
- Distribution: applications are published, licensed and updated through the Univedo Store.
- Origin: built for manufacturing, where it runs production-critical software today; not limited to one industry.
- Case study: EVOMECS runs its complete machining production suite (ERP link, planning, tool management, machine connection, automated cells) on univedo, with more than ten applications on one shared picture.

## Pages

- [Home](https://univedo.com/): what univedo does, the four promises, deployment options, the case study in brief
- [Platform](https://univedo.com/platform/): shared data, cross-system processes, safe onboarding of applications, the Univedo Store, operations
- [Prompted applications](https://univedo.com/prompted-applications/): how AI-generated applications are brought safely into a sensitive production environment
- [Case study: production-critical machining](https://univedo.com/evomecs/): why the EVOMECS case is demanding and how it runs on univedo
- [Services](https://univedo.com/services/): integration projects, applications, review and approval, hosting, AI setup, training
- [FAQ](https://univedo.com/faq/): short answers to the most common questions
- [Contact](https://univedo.com/contact/): walkthroughs on your own data

## Languages

The site is available in English (https://univedo.com/), Deutsch (https://univedo.com/de/), Français (https://univedo.com/fr/), Español (https://univedo.com/es/), Italiano (https://univedo.com/it/), Português (https://univedo.com/pt/), Nederlands (https://univedo.com/nl/). Every page exists in every language under the same path after the language prefix.

## Frequently asked questions

### What is univedo?

univedo is an integration platform for data and processes. It connects the systems, devices and people of an organisation into one shared, versioned picture. Applications from different vendors work on the same data and hand processes to each other, and new applications, including ones generated from a prompt, reach production data only after they have been approved.

### Does univedo replace our ERP, MES or other systems?

No. Every system keeps its job. univedo sits underneath and connects them: each system contributes its part to the shared picture, and processes that cross system boundaries are handed over through the platform instead of through point-to-point interfaces.

### What are prompted applications?

Applications that you describe in plain language and that AI builds for you, on the shared picture of your organisation and in a separate development space with a live preview. Typical examples are shift boards, request forms, inspection checklists, or small portals for customers and suppliers.

### How can an AI-generated application be safe in production?

Through five guarantees. It builds on the shared picture instead of inventing its own data. It is developed in its own space. Every step is recorded as a version. An expert approves it before it may change production data. And the platform checks that approval on every write, so an unapproved application may read but not change what keeps your organisation running.

### Can univedo run on premise or in a private cloud?

Yes. univedo runs in the cloud, on premise on your own servers, or in your private cloud. The on-premise setup is guided, needs no configuration files and only an outbound internet connection. Data never has to leave the site.

### Can we switch between cloud, on premise and private cloud later?

Yes. All three run the same software, so moving between them is a planned migration rather than a new project. Data, applications and approvals move together, and you can also mix: a pilot in the cloud, one site on premise, another in your private cloud.

### Is univedo only for manufacturing?

No. univedo grew up on the shop floor, and its most demanding case today is production-critical machining. The platform itself is not tied to an industry: the shared picture can just as well hold patients and beds, shipments and vehicles, or samples and results.

### Which AI models are used, and does our data leave the site?

You decide. Every AI call from any application goes through one component that you configure: which providers, which models, which monthly limits. Models can run on site when data must not leave your premises. Applications never hold keys to AI services themselves.

### How do applications from different vendors work together?

They all work on the same shared picture and hand processes to each other as events. Each application declares which part of the picture it uses and is approved for it. Vendors build against one platform instead of integrating with every customer landscape.

### What happens to existing applications when the data model changes?

They keep working. Every change to the shared picture is recorded as a new version, and applications built against an earlier version continue to run. That is what lets many vendors extend one platform without breaking each other.

### How are applications bought, installed and updated?

Through the Univedo Store. Vendors publish approved applications and bundles there; customers buy them per machine and per user, in the cloud or as an on-premise licence. Installations fetch what they need from the store and keep it current.

### Where can developers find technical details?

The technical introduction covers the data model, versions, governance, approvals, events, interfaces and client libraries. It is public: read the technical introduction.

### How do we get started?

With a one-hour walkthrough on a sample of your own data: the shared picture, one process across systems and one prompted application built in front of you. From there we scope which systems to connect first and which applications are missing. Book a walkthrough.

## Technical documentation

- [Technical introduction](https://uspec.univedo.com/docs): data model, versions, governance, approvals and signing, events, interfaces, product components, client libraries

## Optional

- [Univedo Store](https://store.univedo.com): applications, bundles and licences