Sub-processors and AI data
This page names the third parties (sub-processors) Ovidens uses to process customer data for its AI and core hosting, and where that processing happens. It is informational and describes the service as it works today; it is not legal advice. The list below covers AI and core data processing.
AI sub-processors
| Sub-processor | Purpose | Processing location | Trains on your API data? |
|---|---|---|---|
| Anthropic (Claude models) | All Vera inference using the workspace's configured model, through the direct Anthropic API | Not limited to the EU | No, by default under Anthropic's commercial policy |
Ovidens does not train models on your data. See Anthropic's commercial data policy for its default no-training policy and exceptions for explicitly shared feedback.
Which Claude model Vera uses is chosen per workspace. See Choosing the workspace AI model.
Core hosting and data sub-processors
| Sub-processor | Purpose | Region |
|---|---|---|
| AWS (Amazon Web Services) | Application hosting, storage, and database | eu-central-1 (Frankfurt) |
| AWS Cognito | User sign-in and authentication | eu-central-1 (Frankfurt) |
| AWS SES | Transactional email (workspace invitations) | eu-central-1 (Frankfurt) |
| MotherDuck | The managed DuckDB data warehouse that stores and processes your analytics data | Hosted on AWS in the EU |
Data residency and retention
Ovidens' core application hosting, storage and database run on AWS in Frankfurt (eu-central-1). Analytics data is stored and processed by MotherDuck, hosted on AWS in the EU. Data sent to Vera's Anthropic API follows a separate processing boundary: AI processing is not limited to the EU. The API's default inference routing is global; see Anthropic's data-residency documentation and Vera's data handling.
Ovidens' retention follows your agreement. Provider-side retention follows the applicable provider terms; an EU-hosted application does not establish EU-only AI processing or zero provider retention.
No training on your data
Ovidens does not train models on workspace data. Anthropic's commercial API does not use inputs or outputs for training by default, as described in its commercial data policy. Ovidens checks each user's permissions before sending resource data and keeps workspace data isolated from other workspaces.
Data Processing Agreement
A Data Processing Agreement (DPA) is available to customers on request. Contact us at contact@ovidens.com to request a copy. The DPA terms are provided by our legal counsel; this page does not restate them.
This page is a factual, good-faith disclosure and is informational, not legal advice. It does not create any contractual term.
AI transparency
How AI works in Ovidens - the models and provider behind Vera, what data reaches them, the human oversight and controls in place, and how this maps to the EU AI Act. Informational, not legal advice.
Integrations
The two ways other software works with your workspace - an AI assistant acting as a person, or an unattended system acting on its own - and how to choose between them.