AI Content Policy

Last updated: 2026-09-30

NexoresFX is operated by Nexores Commerce LTD. Everything generated on the Service is synthetic audio produced by AI models from a text description. This policy explains the principles we work by, how people stay in control of the automated parts of the Service, and how generated files are labelled so their origin is always clear. It has three parts:

• 1. Ethical framework: what we commit to and what we ask of customers.
• 2. Human oversight: where automation is used and where a person decides.
• 3. Content labelling: what is embedded in every file and how to verify it.

1. Ethical framework

Transparency

Everything generated on NexoresFX is synthetic. We say so on the site, in the files themselves and in our documentation. We never present AI-generated audio as a field recording or a human performance. Part 3 below describes the technical detail.

Respect for people

The Service is for sound effects, ambiences and design elements. It must not be used to imitate the voice, likeness or identity of a real person, to fabricate statements, or to produce audio intended to deceive, harass or defame. Attempts to do so are a breach of our Terms of Service.

Lawful and non-harmful use

Generated audio may not be used for fraud, for unlawful surveillance, to incite violence, or to create content that is illegal where it is produced or used. We cooperate with lawful requests from authorities and we act on credible abuse reports.

Human accountability

Automation never has the last word. Account suspensions, refusals that affect a customer's work and abuse decisions are reviewed by a person. Part 2 below explains how.

Data minimisation

We send only the text prompt and generation settings to our audio providers. We do not share your name, email address or account identifier with them, and we do not use your prompts or files to train models. Full details are in the Privacy Policy.

Fair value for creators

We believe AI tools should expand what sound designers and small studios can do rather than erase their craft. We price per generation, we do not claim ownership of your work, and we encourage customers to credit human collaborators wherever they are involved.

How we choose technology partners

We work only with audio model providers that publish terms covering lawful training data, content safety and data protection. Before adding a provider we review its documentation, its data handling commitments and its acceptable use policy, and we re-check these at least once a year and when a provider materially changes its terms. If a provider no longer meets our standards we stop routing generations to it.

What we ask of customers

• Disclose that audio is AI-generated where your audience, platform or local law expects it. We make this easy by embedding the disclosure in every file.
• Do not strip the embedded metadata in order to misrepresent the origin of a file.
• Do not use the Service to imitate real people or to create deceptive content.
• Report misuse you become aware of to support@nexoresfx.com.

2. Human oversight

NexoresFX uses automated systems to generate audio and to screen requests. This part explains where automation is involved, where people are involved, and how you can reach a person when you need one.

Where automation is used

Decisions that always involve a person

• Suspending, restricting or closing a customer account.
• Responding to a report that generated content was used to deceive, harass or impersonate.
• Responding to a legal request about a customer or their content.
• Changing which prompts are refused.
• Adding or removing an audio engine.

Asking for a review

If a prompt was refused, a generation was charged incorrectly, or your account was restricted, you can ask a person to look at it. Open a ticket from your dashboard or email support@nexoresfx.com with the approximate time, the prompt and what happened. We acknowledge review requests within two business days and aim to resolve them within ten. The outcome and the reasoning are sent to you in writing.

Reporting misuse

Anyone, customer or not, can report content they believe was generated through NexoresFX and misused. Send the file or a link, where you found it and why you are concerned to support@nexoresfx.com. Every file we produce carries an identifier in its metadata which lets us trace it to the generating account when it has not been stripped.

Monitoring, audit and incidents

• Refused prompts, abuse reports and enforcement actions are logged and reviewed on a regular schedule.
• We sample generated output to check engine behaviour and labelling integrity.
• We keep records of oversight decisions for at least 12 months so they can be audited.
• Staff with access to customer prompts and files are limited to those who need it for support, safety or engineering, and are bound by confidentiality obligations.
• If we learn that the Service has produced or distributed content that causes harm, we investigate, contain the issue (for example by disabling an engine or a prompt pattern), notify affected customers where appropriate, and record what we changed. Security incidents involving personal data are handled under our Privacy Policy.

3. Content labelling

All audio produced by NexoresFX is labelled as AI-generated in three places: on the website, inside every downloadable file, and in our records.

Why we label

Transparency rules for AI systems, including Article 50 of the EU Artificial Intelligence Act, require synthetic audio to be marked in a machine-readable format so that it can be detected as artificially generated. Many platforms and broadcasters also require disclosure of synthetic media. Labelling at the file level means the disclosure travels with the audio wherever it goes, and it lets us trace a file back to the account that generated it if it is misused.

What is embedded in each file

Metadata is written into the audio container at the moment of generation, before the file is stored or offered for download. Any tags added by the underlying engine are replaced with ours so the disclosure is consistent.

MP3 files (ID3v2.3 tag)

WAV files (RIFF LIST/INFO chunk)

How to verify the label

The metadata is readable with common tools and most media players and digital audio workstations:

ffprobe -hide_banner my-sound.mp3
exiftool my-sound.wav
python -c "import mutagen; print(mutagen.File('my-sound.mp3').tags.pprint())"

What we ask of you

• Keep the metadata. Do not remove or alter the embedded disclosure in order to present a file as recorded or human-made. Normal editing, mixing and format conversion for your project is fine.
• Disclose where required. If you publish content containing our audio on a platform or in a jurisdiction that requires disclosure of AI-generated material, follow those rules. Where you need wording, "Contains AI-generated sound effects" is sufficient for most contexts.
• Do not use it to deceive. Our audio must not be presented as evidence of a real event or as a real person's voice.

Limits of file-level labelling

Container metadata survives copying and most editing, but it is removed by some streaming platforms and can be stripped deliberately. It is not an inaudible watermark. We are evaluating content credentials standards (such as C2PA) and audio watermarking and will update this policy if we adopt them. Regardless of metadata, we keep a server-side record of every generation, which lets us assist investigations into misuse.

Review and contact

This policy is reviewed at least annually and whenever regulation, our technology or our products change in a way that affects it. Material changes are announced on the Service.

Policy questions: support@nexoresfx.com
Support, reviews and abuse reports: support@nexoresfx.com

Nexores Commerce LTD
2nd Floor College House
17 King Edwards Road
Ruislip, HA4 7AE
United Kingdom