Private IT and AI for hospitals, clinics and life sciences in Switzerland

Hospitals, clinics, research groups and pharma teams run collaboration, research databases and private AI on their own servers or on dedicated machines in Switzerland, with patient and study data kept under their control.

Talk to our teamBook a 30-minute briefing

In short

  • Open-source apps and private AI for hospitals, clinics, research and pharma, on your premises or in Switzerland, operated for sensitive health data under the nLPD.
  • Patient records, study datasets and lab results are sensitive personal data under the nLPD.
  • Clinicians want help summarising reports and drafting letters.
Deployment
Your premises, Pilae Cloud in Zurich, or air-gapped
Data class
Health data, treated as sensitive personal data
Private AI
Open models on your machines, prompts stay there
Record
Every change and administrative access, exportable

The rules this sector asks about first

What Pilae does about each one, and where you can check it.

  1. Swiss nLPD, revised FADP

    A data processing agreement that covers sensitive data, with the hosting location for each app named in the contract.

    ContractThe DPA

  2. Professional secrecy, art. 321 SCC

    Confidentiality obligations binding Pilae and each engineer, written with medical professional secrecy under art. 321 SCC in mind.

    Public pageNamed access, every session recorded

  3. The Human Research Act

    The research database and tools on your premises or in Switzerland, with roles per project and every administrative access in the audit log.

    Public pageThe audit log

Every sector and the rules it asks about first

Health data needs an operator that knows what it is holding

Patient records, study data and lab results are sensitive personal data under the Swiss Federal Act on Data Protection. Where they sit, who can reach them and what is recorded about each access are questions your data protection officer, your ethics committee and your auditors will ask.

  1. On your premises

    Apps run on servers in your own hospital or research building, next to the systems they work with, under your physical and network controls.

  2. Or in Switzerland

    Dedicated machines in Zurich, hosted in ISO 27001-certified datacentres. The contract names Switzerland as the location for your apps, data and backups.

  3. Private AI for clinical and research work

    Staff summarise, draft and search with open models on your own hardware. No prompt, report or dataset goes to a public AI service.

  4. Closed to the internet

    Apps answer only on a private mesh network. One hardened gate on port 443 is the single way in, and you decide who may pass it.

  5. Access your auditors can read

    Each change, approval and administrative session is written to an audit log your security office and data protection officer can export.

  6. Backups at a second site

    Daily encrypted backups, stored offsite in the country your contract names and, on Enterprise, a restore test every month recorded in the console.

Discuss your data classes with an engineer.

Thirty minutes with your IT and data protection teams. We come back with a placement proposal for each app and each class of data.

Talk to our team

  • Go deeperPrivate AIA private ChatGPT for your teams in Switzerland: Open WebUI, LibreChat or Dify, open-weight models in your perimeter, answers grounded in your own documents.
  • CompareWhere each alternative stopsTen alternatives and Pilae, across the thirteen jobs an open-source estate needs done.
  • CheckWhat each plan includesStarter, Business and Enterprise, side by side.

How we start with a hospital or research group

  1. Data classification

    With your IT, data protection and research teams we map each app to the data it will hold: patient data, coded research data, or neither.

  2. Placement decision

    Together we decide what runs on your premises, what runs in Zurich and what stays air-gapped.

  3. Fixed-price onboarding

    We install the apps, connect Keycloak or Entra ID with multi-factor authentication and move your data off the tools you are leaving.

  4. Acceptance

    Your team tests each app, its roles and its access rules before any real data goes in.

  5. Managed operations

    The Pilae Agent plans each update and waits for your approval. Monitoring alerts reach Pilae engineers around the clock.

Sensitive health data, kept where you decide

Patient records, study datasets and lab results are sensitive personal data under the nLPD. The first question is where they sit. Pilae runs your apps on your premises, on dedicated machines in Zurich, or air-gapped, and the contract names each location. Backups follow the same rule: stored offsite in the country your contract names and, on Enterprise, tested every month. The data residency page explains where each copy lives.

Private AI for clinical and research teams

Clinicians want help summarising reports and drafting letters. Researchers want to search papers and protocols. The documents they would paste into a public chatbot are exactly the ones that must not leave. We run private AI on your own hardware, with open models and Open WebUI as the interface. The same setup serves a hospital department or a research group.

Research databases and tools under the HRA

Research under the Human Research Act often works with coded data, and it needs project-level access and a record of who reached what. We operate PostgreSQL for study databases, Metabase for dashboards and Nextcloud for protocols and shared files, with single sign-on through Keycloak and roles per project.

Pharma and medtech teams that answer for every change

A validated system is only as good as its change record. The Pilae Agent plans every update, waits for your approval, applies it in your window after a verified backup and records each step. Your quality team gets the plan, the approval and the result for every change, exportable to your own systems. The security page explains how access and logging work.

Talk to us

Contact us to discuss your data classes and constraints.

What your contract includes

Written into your contract with Pilae, under Swiss law.

  1. Data location

    Named per app: your premises or Switzerland. Apps, data and backups stay there unless you agree otherwise in writing.

  2. Data processing

    A data processing agreement under the nLPD, with the sub-processor list and technical measures for sensitive data.

  3. Access

    Named Pilae engineers, over the private network, under access rules you set. Every session is recorded.

  4. Confidentiality

    Confidentiality obligations binding Pilae and each engineer, written with medical professional secrecy under art. 321 SCC in mind.

  5. Availability

    Monthly availability commitment with service credits: 99.5% on Business, 99.9% on Enterprise.

  6. Exit

    A full export of data and configuration in open formats, at any time and at the end of the contract.

Where it depends on the plan

ClauseStarterBusinessEnterprise
Availability commitmentNone. Monitored around the clock, without service credits.99.5% a month, with service credits.99.9% a month, with service credits, or negotiated.

Who runs it

Same platform, different obligations

The rules each sector asks us about first. The platform underneath is the same for all of them.

SectorSwiss nLPDRevised FADPGDPRRegulation (EU) 2016/679FINMACircular 2018/3DORAFor EU entitiesProcurement lawWTO GPA, BöB, IVöBProfessional secrecyArt. 321 SCCCLOUD ActUS reachThe sector’s own
Public sectorSwiss nLPDSwitzerland onlyNot among its first questionsNot among its first questionsNot among its first questionsProcurement law: WTO GPA, BöB/LMP, IVöB/AIMPTendersNot among its first questionsNot among its first questionsEMBAG, the federal open-source actEMBAG
International organisationsNot among its first questionsNot among its first questionsNot among its first questionsNot among its first questionsNot among its first questionsNot among its first questionsThe US CLOUD ActWhose law can reachHeadquarters agreement and the Host State ActHost State Act
SportsSwiss nLPDAthlete health dataNot among its first questionsNot among its first questionsNot among its first questionsNot among its first questionsNot among its first questionsNot among its first questionsNot among its first questions
FinanceNot among its first questionsNot among its first questionsFINMA Circular 2018/3OutsourcingDORAEU entitiesNot among its first questionsNot among its first questionsNot among its first questionsBanking secrecy, art. 47 of the Banking ActBanking secrecy
Health and life sciencesSwiss nLPDSensitive dataNot among its first questionsNot among its first questionsNot among its first questionsNot among its first questionsProfessional secrecyMedical secrecyNot among its first questionsThe Human Research ActHuman Research Act
Legal and fiduciarySwiss nLPDClient dataNot among its first questionsNot among its first questionsNot among its first questionsNot among its first questionsProfessional secrecyLawyers, notariesThe US CLOUD ActClient filesNot among its first questions
Research and educationSwiss nLPDProcessor DPAGDPRWhere it appliesNot among its first questionsNot among its first questionsNot among its first questionsNot among its first questionsNot among its first questionsFunder data management plansFunders’ data plans

Questions

Is health data treated differently under the nLPD?

Yes. The Swiss Federal Act on Data Protection classes data on health as sensitive personal data, which raises the bar for consent, disclosure and security measures. Your contract includes a data processing agreement that covers sensitive data and the hosting location for each app. Public hospitals usually fall under cantonal data protection law instead, and cantonal health laws may add rules. We work with your data protection officer on whichever applies.

Can research data under the Human Research Act run on Pilae?

Yes. Research under the Human Research Act (HRA) often works with coded data and needs access limited to the project team and a record of who reached the data. We run the research database and tools on your premises or in Switzerland, with roles per project and every administrative access in the audit log. The approval of your ethics committee and the coding of the data remain with your research team.

Can clinicians use AI without sending patient data to a public service?

Yes. Open WebUI runs on your machines with open models hosted there. Prompts, uploaded documents and answers stay inside your environment. The optional connection to an external model through an AI gateway stays off unless you decide to enable it.

Where are backups of patient and study data kept?

Backups run daily, are encrypted and are stored offsite in the country your contract names, which for health data is usually Switzerland. On the Enterprise plan, a restore test runs every month and its result is recorded in the console. Retention follows your plan, and on Enterprise the recovery point and recovery time objectives are written in the contract.

Does Pilae replace our hospital information system?

No. We operate the open-source apps around it: file sharing, internal documentation, research databases, dashboards, workflow automation and private AI. They connect to your existing systems through their standard interfaces, under access rules you set.

Does Pilae hold a health-sector certification?

No. Our operations are built to ISO 27001 controls, and Pilae Cloud machines are hosted in ISO 27001-certified datacentres. For a pharma team validating a system, the audit log gives a complete change record: plan, approval, result and rollback for every change.

Discuss your data classes with an engineer.

Thirty minutes with your IT and data protection teams. We come back with a placement proposal for each app and each class of data.