Enterprise AI platform: featuresCollapps in detail: build, publish and govern your AI business applications
Who builds, who approves, how an application moves from prototype to production, and how AI usage stays under control.
Enterprise AI application platformWhat is Collapps?
Hi-perf Collapps is a platform for industrialising internal business applications and their AI usage. Selected key users create or improve tools based on needs on the ground; the IT department keeps control of access, code, environments, releases to production and AI usage.
Collapps is not a platform where every employee becomes a developer. Nor is it a catalogue of generic applications. It is a framework: it lets the people who know the processes build the tools they need, without creating a new form of shadow IT.
Collapps runs as SaaS, hosted by Hi-perf, or installed on a machine in your company: you choose where the platform, your applications and their data live.
Key users, IT department, usersThree roles, one shared framework
Key users
Selected, trained and authorised, they know the pain points of their job: double data entry, manual checks, intermediate files, blind spots in the business software. Assisted by AI, they create or improve applications within the framework defined with the IT department.
The IT department
It defines that framework: who can build, which AI models are approved, which checks apply, who signs off releases to production. It keeps an overview of all applications and their usage.
Users
They access the applications published in the internal portal, according to their permissions. For them, a Collapps application is simply a company tool.
From AI prototype to productionThe lifecycle of an application
- Business need
- Key user
- Assisted development
- Git
- Checks
- Orchestrator
- Test
- Approval
- Production
- Internal portal
Git. Each application’s code is versioned in a repository. The company retains ownership of it, and the history of changes is kept.
Checks and technical compliance. Before publication, the application is checked and, if necessary, brought into line with the defined technical standards: structure, prerequisites, security.
Orchestrator and environments. The application is first deployed to a test environment, then to production after approval. Building quickly with AI is no excuse to skip TEST and PROD: that is what makes the tool usable over time.
Internal portal. Published applications are gathered in a library in the company’s branding, with per-application permissions.
Approved models, quotas, token costsAI governance
Each application can call on AI, within a defined framework:
- The company’s AI accounts: you use your own contracts with providers (OpenAI, Anthropic, Google, others), or your own open-source model hosted on your premises.
- Approved providers and models, chosen by the IT department.
- Usage tracking, per application and per model.
- Quotas, caps per application or per period, and alerts, up to blocking beyond a threshold.
Collapps does not force a model provider or token purchases on you: it provides governance, tracking, caps and industrialisation.
Usage data is not just for keeping the budget in check: it also helps spot a poorly designed application or abnormal usage, choose between models and compare cost with value. AI becomes a managed service, not a diffuse expense that cannot be attributed.
Specific option · quote on request
Sovereign AI: on-premise open-source LLMYour own AI, built on an open-source model
An alternative to consuming tokens on commercial AI services: Hi-perf helps you build a proprietary AI from an open-source model, installed on your infrastructure and used by your Collapps applications like any other approved model.
- What you can doHost an open-source model on your premises, adapt it to your uses as part of the project, and offer it to your applications alongside, or instead of, commercial AI services.
- What you need to plan forA GPU-equipped machine, purchased by your company. Hi-perf provides hardware recommendations suited to your uses and volumes.
Advantages
- Your data never leaves your infrastructure.
- No per-token billing: a predictable cost, tied to the hardware.
- Independence from AI providers.
- Control over the model and how it evolves.
Drawbacks
- A hardware investment and its running costs: energy, maintenance.
- On some tasks, open-source models still lag behind the most advanced commercial AI.
- Sizing to plan according to your uses and volumes.
- A specific project, with lead time to set up.
The two approaches can be combined: a model hosted on your premises for sensitive data or the most frequent uses, commercial AI for the rest. Collapps applies the same governance to all of them.
SSO, permissions, Git, portal, AI monitoringFeatures, by area
- GovernanceSSO, users, roles, per-application permissions, access control.
- DeploymentGit, test and production environments, approval, orchestrator, prerequisite checks.
- AIApproved models, company accounts, quotas, caps, usage tracking.
- PortalApplications available according to permissions, centralised access, favourites, notifications.
- MethodSelected and trained key users, development framework and rules, Hi-perf support.
Anonymise a document before AIExample: anonymisation before analysis
To show what a governed AI application changes, take a simple case: having a document containing personal data analysed. A Collapps application chains anonymisation, checking, the model call and delivery of the result. AI is no longer called directly: it fits into a controlled process, with rules that apply even before the model is called. Data detection and masking are configurable; for sensitive documents, human review may still be needed.
Individual use, no-code or governed AI platformCollapps compared with other approaches
| Individual AI use | No-code without a framework | Traditional IT development | Collapps | |
|---|---|---|---|---|
| Who builds | Every employee | Any user | IT department teams | Selected key users |
| Speed | Immediate | Fast | Depends on the backlog | Fast, AI-assisted |
| Knowledge retention | None | Low, scattered tools | High | High: versioned code, shared portal |
| IT control | None | Limited | Full | Controlled framework, approval and releases |
| AI costs | Invisible | Barely tracked | — | Measured, capped per application |
Enterprise AI platform pricingPricing
Collapps starts at €500 excl. VAT per month. Your price depends on the number of users, the level of support you want and the hosting option; a setup is mandatory for an installation on your premises. AI usage stays on your own accounts; a proprietary AI on an open-source model is an option priced on quotation. See pricing details.
Enterprise AI platformFrequently asked questions
Do you need to know how to code to build an application?
Development is AI-assisted, which greatly lowers the technical barrier. But Collapps is aimed at selected, trained key users, not at every employee indiscriminately: that is a condition for quality and governance.
Where is Collapps hosted?
Your choice: as SaaS, hosted by Hi-perf, or installed on a machine in your company. In the latter case, the platform, the applications and their data stay within your environment, and a setup is mandatory.
Which AI providers can we use?
Those your company chooses and approves, with its own accounts. The IT department defines which providers and models applications may use. As an option, you can also have your own open-source model, hosted on a GPU-equipped machine.
What happens to an application if its creator leaves the company?
It stays. Its code is versioned in Git, it is deployed by the orchestrator and published in the internal portal: it belongs to the company, not to an individual.
Does Collapps replace our business software?
No. Collapps applications address the pain points around your software: double data entry, manual checks, repetitive tasks. Your core tools stay in place.
Turn individual optimisations into company assets.
Let’s talk about your key users, your IT department and the first pain points to tackle.