AIGC Use Policy

Scope

This policy applies to stories, names, prompts, reference-song choices, lyric edits, song and cover directions, generated lyrics, songs, covers, and published Song Gift pages created with Dearlume. It forms part of our Terms of Service.

You are responsible for the material you submit and for how you use and share generated results. A reference song may guide high-level traits such as mood, tempo, instrumentation, vocal character, and production feel. It must not be used to copy protected lyrics, melody, recordings, an artist's identity, or a living person's voice.

AI Models and Generation Services We Use

The current Song Gift generation path uses:

• Lyrics: Suno lyrics-generation service.
• Song candidates: Suno V6.
• Keepsake cover: Nano Banana 2.

Dearlume may connect to these models and generation services through third-party API infrastructure. This disclosure identifies the models and generation services used in the product; it does not imply that Dearlume integrates with them directly or has an official partnership with their providers.

Reference-song analysis and Creative Canon compilation may use Kie-hosted DeepSeek, OpenAI, or Anthropic according to our service configuration. If reference-song analysis is unavailable after a permitted request, Dearlume shows that it has fallen back to the selected music-style preset instead. The current US$9.99 Song Gift does not include video generation. Model availability can change, and this page will be updated when a material provider or model change affects the product.

Dearlume is an independent service and does not imply sponsorship or endorsement by these providers.

Content Standards and Prohibited Categories

Do not request, upload, generate, publish, or share a Song Gift containing or facilitating any of the following:

1. Pornographic and NSFW content. Sexually explicit material, sexual services, non-consensual intimate imagery, or sexualized nudity.
2. Violence and graphic gore. Content that glorifies, instructs, or graphically depicts severe injury, torture, mutilation, or real-world violent wrongdoing.
3. Hate speech. Content that attacks, dehumanizes, threatens, or promotes discrimination against people based on protected characteristics.
4. Child sexual abuse material and other child-unsafe content. Any sexualization, grooming, exploitation, or endangerment of a person under 18 is strictly prohibited. Do not upload suspected CSAM when reporting it.
5. Deepfakes and impersonation. Deceptive use of another person's identity, likeness, or voice; voice cloning without authorization; or content falsely presented as an authentic statement or performance by a real person.
6. Copyright and trademark infringement. Unauthorized copying of protected lyrics, melodies, recordings, characters, logos, brands, or distinctive artistic expression; requests to imitate a living artist or mislead people about affiliation or endorsement.

Content that is illegal, fraudulent, privacy-invasive, harassing, or otherwise harmful is also prohibited even when it does not fit neatly into one category above.

How We Moderate Content

Dearlume combines several controls:

• Before sending a text request to an AI generation service, Dearlume uses Waffo Prompt Sift as an automated text-prompt safety check. Generation continues only when the check returns an explicit allow. A block, review result, degraded service, unavailable check, or unconfirmed response stops that request.
• Text newly generated through the current lyrics, reference-song analysis, and Creative Canon workflows is checked before Dearlume displays or saves it. Final lyrics are checked again when the creator confirms them and immediately before publication. Older lyrics drafts without the current moderation marker are withheld unless the current workflow successfully re-evaluates them.
• Local request validation also blocks detected attempts to imitate living artists, clone voices, or use protected characters and brands improperly.
• AI providers may apply their own safety controls to prompts and generated results. Provider refusals are treated as a failed or refused generation, not as successful delivery.
• Published content and reported content may be reviewed against this policy, the Terms of Service, provider rules, and applicable law.

Prompt Sift checks text. It does not establish that reference images, generated covers, audio, or video have been independently classified by Waffo. Provider controls and creator review therefore remain important. These controls reduce risk but are not perfect and do not mean that every result has been manually approved. Users must review a generated result before publishing or sharing it.

Reporting

Anyone, including a recipient who is not signed in, can report a suspected violation by emailing support@dearlumegifts.com or using the Report inappropriate content link on a public gift page. Signed-in users can also use Settings → Support Tickets. Include the public gift link, the type of suspected violation, and enough context for us to locate and assess it. Do not include passwords, full payment credentials, or illegal material in the report.

We will acknowledge content reports within 2 business days and complete an initial review within 5 business days. If an investigation needs longer, we will update the reporter within 5 business days with the next steps and expected timing. We prioritize credible reports involving child safety, threats, impersonation, or unlawful content and take appropriate action after review.

Enforcement

Depending on the severity, context, and available operational controls, Dearlume may:

• refuse or stop a generation request;
• investigate a reported public gift and take available access-control or account action; or
• preserve relevant records and cooperate with providers or lawful authorities where required.

Some unpublish, account-restriction, and case-audit controls still require additional operator tooling. We will not represent a report as resolved until the available action has actually been taken. A signed-in user may ask for a review of an enforcement decision through Settings → Support Tickets.

Questions

Questions about this policy can be sent to support@dearlumegifts.com; signed-in users can also use Settings → Support Tickets. This policy does not replace the Terms of Service, Privacy Policy, provider terms, or applicable law.