Your knowledge. Your choice of processing.
Choose the processing configuration appropriate to your content. Enterprise offers supported worldwide provider options, EU-based processing or internal processing on KnowScapes-operated infrastructure in Germany.
Personal workload content: Enterprise internal Germany only.
Three Enterprise processing choices
| Configuration | AI workload processing | Personal workload content |
|---|---|---|
| Supported worldwide provider | Selected external-provider configuration; its terms apply | Not permitted |
| EU processing | Configured EU endpoints and routing for ingestion and inference | Not permitted |
| Internal Germany | KnowScapes-operated infrastructure in Germany, without external AI processors | Permitted subject to the applicable data processing agreement |
The selected processing boundary covers ingestion and AI inference, including routing and fallbacks. EU processing means processing inside the EU; a provider's headquarters alone do not establish it. Availability of individual models and providers depends on the supported configuration.
Internal Germany processing is exclusive to Enterprise. It is not a Professional or Business add-on and does not imply dedicated hardware, offline operation or an arbitrary country choice.
What stays within internal Germany processing
In this mode, source content, prompts, retrieved context and answers are not sent to external AI processors. Both ingestion and inference use KnowScapes-operated infrastructure in Germany. Fallbacks must preserve that boundary.
This statement concerns AI workloads. Original source-file storage, account administration and backups are separate. Connecting a Google Drive folder, for example, does not move the original files to Germany.
Content rules apply to every intake path
Community and every external AI configuration require non-personal workload content. This includes websites, uploaded files, connected documents, page context, conversation messages and prompts. Your own provider API key does not remove this restriction. Neither does choosing an external provider's EU endpoint.
Personal workload content is permitted only in Enterprise internal Germany mode under the applicable data processing agreement. Account and billing data are a separate category described in the service privacy notice.
Hosting and AI processing are different
The service description identifies EU/EEA hosting and storage infrastructure in Germany and Finland. AI processing may take place elsewhere when an external provider is selected. Internal Germany AI workloads exclude the Finland locations.
Assess the selected AI configuration and the original source-storage arrangements separately. A single regional label does not describe every data category.
What KnowScapes retains
Original files are processed and are not retained as stored source documents. Their learned knowledge remains in the Knowledge Graph. Processing metadata supports document identity and capacity counting.
Removing a source file, its source record or its connection does not erase learned knowledge and does not free a document slot. Updating a source does not guarantee selective erasure of earlier facts.
How knowledge deletion works
Individual document contributions cannot be selectively removed. The Account Owner can permanently delete the entire Knowledge Base, removing its active graph. Backup copies expire through the regular backup cycle; deleting the active graph is not a promise that every physical backup copy disappears immediately.
If you need to remove particular source-derived knowledge, source deletion alone is insufficient. Plan your Knowledge Base scope with this lifecycle in mind.
Choose a boundary before adding sensitive content
Confirm your selected mode, document sources and applicable data processing agreement before submitting personal workload content. Other service terms and provider-specific processing conditions remain applicable.
Choose the processing boundary first.
Then add sources appropriate to that boundary and keep original source storage and knowledge retention in view.