Videola

AI Transparency Notice

Last updated: 28.07.2026

This notice explains how Videola uses artificial intelligence, how generated videos are marked, the limitations of those markings, and what you should consider when publishing generated content. Videola has implemented transparency and provenance measures intended to support compliance with Article 50 of Regulation (EU) 2024/1689 (the EU AI Act), whose relevant obligations apply from 2 August 2026. It should be read together with our Terms of Service and Privacy Policy.

1. How Videola uses artificial intelligence

Videola turns narration text and other selections into an assembled video. Depending on the features you select, Videola may use artificial intelligence to interpret, translate, or transform narration text; analyse the context of a script; divide narration into scenes or visual segments; create image-generation prompts; generate images; produce synthetic voice narration; select or generate background music, from an AI-generated catalog or a track you upload; generate or synchronise captions; and make automated creative decisions concerning imagery and scene selection.

Videola currently uses OpenAI for text-related processing and AI image generation, and ElevenLabs for synthetic voice generation. Background music is either selected from Videola's AI-generated music catalog (created with Suno) or uploaded by you. When Beast Mode is selected, relevant portions of your script may be analysed to identify scenes, create contextual image prompts, and generate corresponding imagery.

2. Uploaded images

Images you upload are not sent to OpenAI, ElevenLabs, or any other AI-model provider. They are temporarily stored in Videola's private Cloudflare R2 bucket under its European Union jurisdiction configuration, retrieved by Videola's server-side infrastructure, and processed with FFmpeg solely to incorporate them into your video. Uploaded images are deleted after the generation job completes or fails; an automated cleanup process also removes files left behind by interrupted, abandoned, or failed jobs within 48 hours. Further information is in our Privacy Policy.

3. Uploaded music

If you upload your own background music track, it is not sent to OpenAI, ElevenLabs, or any other AI-model provider. It is temporarily stored in Videola's private Cloudflare R2 bucket under its European Union jurisdiction configuration, retrieved by Videola's server-side infrastructure, and processed with FFmpeg solely to mix it into your video. Uploaded music is deleted after the generation job completes or fails; an automated cleanup process also removes files left behind by interrupted, abandoned, or failed jobs within 48 hours. Further information is in our Privacy Policy.

This is separate from Videola's own AI-generated music catalog (created with Suno), which is not user content and is not affected by this cleanup process.

4. Which parts of a Videola video are AI-generated

A video's imagery may consist of entirely AI-generated images, entirely user-uploaded images, or a combination of both. Background music, where included, is either an AI-generated track from Videola's catalog or a track you uploaded yourself. Captions, timing, transitions, effects, and other elements may be generated or assembled automatically. Every Videola-generated video is therefore an AI-assisted production — though this does not mean every visual, audio, or other element was itself generated by AI.

5. How Videola informs you about AI use (in-app transparency)

Videola shows AI-use notices to the person creating the video at the following points in the Service:

  • an EU AI Act transparency card in the Voice step;
  • a source label ("AI" or "Uploaded") on each image in the Images step;
  • a persistent AI-generation notice in the Review step, shown at the point you commit to generating;
  • an "AI-generated content" indicator on the result screen, together with a publishing helper (a copy-ready disclosure line and platform guidance); and
  • an "AI-generated content" indicator in your video history.

These in-app notices inform you, the creator. They do not themselves mark the exported file for downstream detection, and they do not replace any visible disclosure that may be required when the video is published (see Section 9).

6. Machine-readable marking of the file

Videola embeds a C2PA Content Credentials manifest in the MP4 videos it generates. C2PA Content Credentials provide machine-readable provenance information cryptographically bound to the exported file, making it possible for compatible tools and platforms to detect that the video was generated or assembled using artificial intelligence. This implementation is intended to support Videola's obligations under Article 50(2) of the EU AI Act, which requires relevant AI-generated or manipulated content to be marked in a machine-readable manner using measures that are effective, interoperable, robust, and reliable as far as technically feasible.

C2PA is Videola's current provenance mechanism. Each finished video shows its Content Credentials status on the result screen so you can confirm the manifest was embedded. The presence of Content Credentials does not guarantee that every application or publishing platform will recognise, preserve, or display them (see Sections 7, 8, and 10).

Additional metadata. In addition to the C2PA manifest, Videola writes a container metadata field in each exported MP4 — the comment field, set to: "AI-generated with Videola: synthetic voice and AI-generated imagery. Disclosed under EU AI Act Article 50." This is a plain machine-readable string; the C2PA manifest is the cryptographically signed, tamper-evident layer over the same file.

7. Information contained in the C2PA manifest

Videola's C2PA manifest identifies Videola as the service that generated or assembled the file. Depending on the video, it may record: the Videola service name and identifier; a generic title identifying it as a Videola-generated video (not derived from your script or personal to you); that AI-generated or AI-modified components were used; the creation action; the type of digital source used (trainedAlgorithmicMedia for fully AI-generated imagery, or compositeWithTrainedAlgorithmicMedia where AI and uploaded material are combined); a disclosure statement matching the metadata field above; a processing timestamp; and cryptographic information used to bind the manifest to the video.

The C2PA manifest is not intended to identify you. Videola does not intentionally include your name, email address, account identifier, payment information, narration text, prompts, uploaded images or filenames, private storage addresses, or internal project/job identifiers in publicly readable C2PA claims.

8. Current signing certificate

Videola currently signs its C2PA manifests using a self-signed certificate. This means compatible tools may be able to inspect the manifest and verify its cryptographic binding to the file, but may not recognise Videola's signing identity as trusted. A verification service may display a message such as "unrecognised signer", "unknown signer", or "untrusted signing credential". This does not mean the manifest is absent — it means the certificate does not currently chain to a trust anchor recognised by that service. Videola may adopt a certificate recognised through the C2PA Trust List or another recognised trust mechanism in future. Because of this, we do not promise that any particular platform will recognise or display Videola's Content Credentials.

9. Limitations of Content Credentials

Content Credentials provide provenance information and evidence of whether the credential has remained cryptographically bound to a file. They do not prove that the narration or imagery is factually accurate, that the events shown occurred, that an uploaded image is authentic, that you own or have permission to use all included material, that the video complies with applicable law, or that it is free from defamatory, misleading, or infringing content. A cryptographically signed synthetic video can still contain false information.

Content Credentials may be removed, become inaccessible, or fail validation when a file is edited, re-encoded, compressed, converted, modified after export, processed by software that does not preserve C2PA information, or uploaded to a platform that removes or changes metadata. A broken or missing credential does not prove a video was not created with Videola; equally, the presence of a credential does not prove the truthfulness of the video's claims.

10. Your responsibilities when publishing

You are responsible for reviewing a completed video before publishing or distributing it, and for assessing whether your intended publication requires a visible disclosure under applicable law, advertising or consumer-protection rules, election or political-content rules, the policies of the platform you publish on, or standards applicable to you. Videola's machine-readable Content Credentials do not automatically satisfy a visible-disclosure requirement imposed on the publisher.

Professional and commercial use in the EU. A person or organisation using an AI system under its authority in a business, trade, professional, occupational, or regular monetised context may qualify as a "deployer" under the EU AI Act. A natural person using AI solely in the course of a personal, non-professional activity is generally outside the AI Act's deployer obligations, though other laws and platform policies may still apply. Where you qualify as a deployer, Article 50 may require you to disclose AI-generated or manipulated image, audio, or video content that constitutes a deepfake, and certain AI-generated or manipulated text published to inform the public on matters of public interest where there has been no human review or editorial responsibility. A deepfake generally means AI-generated or manipulated content resembling an existing person, object, place, entity, or event that would falsely appear authentic or truthful. Where a visible disclosure is required, it should be clear and perceivable when the audience first encounters the content.

11. Publishing on YouTube and other platforms

Publishing platforms maintain their own disclosure rules, which may change independently of Videola. YouTube currently requires creators to disclose meaningfully altered or synthetic content when it appears realistic — for example, content that makes a real person appear to say or do something they did not, alters footage of a real event or place, or creates a realistic-looking scene that did not occur. Clearly unrealistic material, minor aesthetic edits, and some production assistance may not require it. Where the rule applies, use YouTube's current altered-or-synthetic-content disclosure setting, or any replacement tool YouTube provides. Other platforms may apply different standards, and you are responsible for checking the current rules of every platform on which you publish. The presence of Videola's Content Credentials does not guarantee that a platform will automatically display an AI label.

12. Suggested visible-disclosure text

When a video is completed, Videola may provide a ready-to-use disclosure statement. A general disclosure may state:

This video was created with AI-assisted tools and contains synthetic voice narration and/or AI-generated imagery.

For realistic content depicting a real person, place, object, or event, a more specific disclosure may be appropriate:

This video contains artificially generated or manipulated audio or imagery and should not be understood as an authentic recording of the depicted event.

Using a disclosure does not make otherwise unlawful, defamatory, deceptive, or infringing content lawful.

13. Real people, voices, and events

You must not use Videola to falsely represent that a real person said or did something; depict a real person, object, place, entity, or event in a deceptive manner likely to cause harm; create non-consensual intimate content; exploit or sexualise minors; commit fraud, impersonation, or identity theft; conceal legally required AI disclosures; or remove or alter provenance information to deceive others. Having consent or adding an AI disclosure does not override another applicable law or platform prohibition. Additional restrictions are stated in our Terms of Service.

14. AI-generated content can contain errors

AI-generated narration, imagery, prompts, captions, and translations may contain factual inaccuracies, mistranslations, mispronunciations, visual defects, inappropriate or unexpected elements, misleading representations, or similarities to existing material. You must review every video before publishing it. Extra care is required for content concerning health, finance, legal matters, elections, public safety, conflicts, disasters, or other subjects where inaccurate content could cause significant harm.

15. AI providers and privacy

Videola currently uses OpenAI for text-related processing, contextual analysis, and image generation, and ElevenLabs for voice synthesis. Uploaded images are processed with FFmpeg on Videola's server-side infrastructure and are not submitted to OpenAI or ElevenLabs. The personal-data and retention implications are described in our Privacy Policy.

16. Changes to this notice

We may update this notice when our AI providers change, when we introduce new generation features, when our C2PA implementation or signing certificate changes, when publishing-platform requirements change, or when applicable law or regulatory guidance changes. The date at the top identifies the latest version. For material changes affecting how Videola marks generated content or informs you about AI use, we will provide appropriate notice through the Service.

17. Contact

Questions about Videola's use of artificial intelligence or the provenance information attached to a video:

Oleg Tatarciuc
Self-employed professional / autónomo
NIF/NIE: Z1704146N
Email: support@videola.ai\