ZahlenwerkDocs

Architecture & data storage

Where Zahlenwerk runs, where your data is stored and which service providers are involved.

Zahlenwerk is a web application operated entirely by Zahlenwerk (software as a service). You install nothing; updates roll out continuously without maintenance windows.

Components

ComponentProviderRegion
Application and APIVercelDublin (EU)
Database, sign-in, file storageSupabase (AWS)Ireland (EU)
Background processing (extracting reports)runs in the Zahlenwerk databaseIreland (EU)
AI modelsGoogle Vertex AI (Gemini)EU (Frankfurt / Netherlands)
PDF conversionModalEU
Transactional emailsPostmarkUSA

AI processing in the EU

  • Without your own key, AI models are called only in EU regions. If EU processing can't be guaranteed, the call fails instead of falling back to another region.
  • When documents are extracted, names of people and entities from your directory are replaced with placeholders before the AI call and only put back in the result.
  • Your data is not used to train AI models.

Tenant separation

Each team is its own tenant. The separation is enforced in the database itself (Row Level Security), not just in the interface. AI clients and background jobs, too, see only the data of the team in question.

Subprocessors

The binding list is in the data processing agreement (AVV). Beyond the ones above, services are only used if your team actively connects them, such as Microsoft (SharePoint, Teams), Slack or Google Drive, or if you add your own AI key.

More information: zahlenwerk.studio/security.

On this page

EN
EnglishDeutsch