Subprocessors
This page lists third-party service providers ("subprocessors") that VidiLex uses to deliver the service.
We keep this list intentionally small and purpose-limited.
Summary
- • Subprocessors are used only where required to operate core features
- • Data access is purpose-limited and contractually restricted
- • Customer data is not used to train models
What is a subprocessor?
A subprocessor is a third party that processes data on behalf of VidiLex to provide a specific function (for example transcription, AI inference, or email delivery).
Our principles
Data minimization (only what's needed)
Purpose limitation (only to deliver VidiLex)
Contractual safeguards (data processing agreements)
Tenant isolation (no cross-customer access)
Transparency (this list is maintained and reviewed)
Current subprocessors
| Subprocessor | Function | When used | Data involved | Processing location | Compliance | Notes |
|---|---|---|---|---|---|---|
| Netcup GmbH | Server and storage | Continuously for hosting and storage | All customer data, video files, transcripts, metadata, search indexes, and user data | Germany | View compliance | |
| AssemblyAI | Transcription | During ingestion of uploaded videos (users can choose between US and EU endpoints) | Audio/video content for transcription, timestamps, detected language | Ireland / USA | View compliance | All data is immediately deleted after transcription. |
| Nebius AI | Embeddings + EU endpoint LLM operations | During ingestion (embeddings) and when EU endpoint is selected for LLM operations | Text segments for embeddings and LLM requests | Netherlands | View compliance | Zero-day retention policy. |
| OpenRouter | US endpoint LLM operations | When US endpoint is configured for LLM operations | Text segments for LLM requests | USA | View compliance | Zero-day retention policy. |
| Resend | Email invitations | When sending user invitations | Recipient email address and invitation content | USA | View compliance | Currently in review |
Notes on endpoint configuration
VidiLex supports configurable endpoint routing for some AI operations:
- • EU endpoint: LLM operations use Nebius, transcript services use eu.assemblyai.com
- • US endpoint: LLM operations use OpenRouter, transcript services use assemblyai.com
- • Embeddings use Nebius regardless of endpoint setting
(See application configuration and provider routing documentation.)
Changes to subprocessors
If we add or materially change subprocessors, we will update this page and provide appropriate notice for customers where required.
Questions?
If you need additional documentation for legal or procurement review, contact us.