Vauda
Legal

Data & AI Policy

Last updated: July 24, 2026

The most important question you can ask an AI product is: does my material train any model? At Vauda the answer is no. This policy explains exactly what the AI does with your content, what we never do, and what control you have.

01The principle

Vauda uses artificial intelligence to catalog your library —not the other way around—. Your content works for you: it is processed to make it searchable and organized, and it never becomes training material for any model, ours or a third party's.

02What the AI does with your content

When you upload material, we process it with AI solely to catalog it and make it searchable for you. Specifically, it:

  • Generates descriptions and tags of what appears in your images, videos and documents.
  • Transcribes the audio of your audio and video files, with timestamps and speakers.
  • Detects and extracts visible text (OCR) and the text of your documents.
  • Generates vectors (embeddings) that enable search by meaning within your own library.

All of this is done on your behalf and for your organization. The resulting catalog is yours, and only you (and those you authorize) can see it.

03What we never do

  • We do not use your content to train, fine-tune or improve AI models —neither ours nor third parties'—.
  • We do not sell or share your content for advertising.
  • We do not use one organization's material to improve another's experience.

04Our AI providers

To catalog, transcribe and generate the search vectors we work with specialized AI providers. They all process your content solely on our behalf, under contract, and none of them uses it to train their models: where a provider trains by default, we have explicitly activated the training opt-out.

05Automated processing and your control

The catalog is generated automatically, is an aid and may contain errors. You keep full control: you can edit any field, add your own information (titles, notes, tags), reprocess an asset or delete it. Cataloging does not make decisions with legal effects or evaluations about people: it only describes and organizes your material.

06How we describe people

When people appear in your material, we instruct the AI to describe them respectfully —by their role, their action, their expression or the context— and to avoid sensitive or offensive characterizations. The goal is a useful, dignified catalog, not a judgment about anyone.

07Sensitive content and responsibility

Because your library may contain images, voices or data of identifiable people, when you upload that material you represent that you have the necessary rights and authorizations. Regarding that content, you (or your organization) are the controller and we act as processor, following your instructions.

08Retention and deletion

The content and the derived catalog —descriptions, transcriptions, detected text and search vectors— are deleted when you delete the asset or your account. We keep no versions of your content inside any model, because it never enters one.

09Questions

If you have questions about how we process your content with AI, write to us at [email protected].

We publish every provider that touches your material, what exactly it touches and its public commitment, with the link to verify it. See it on our Security page.