AI-Generated Content Policy

1. Scope and AI-generated disclosure

This policy applies to the following content that is generated or edited by AI on the service. All of it is AI-generated, not created by a person.

• Text: reading text from each tool (BaZi readings, annual reports, and similar) and consultation answers.
• Audio and music: lyrics and songs generated from a chart (chart songs).
• Images: the cover image that the music model produces together with a song.
• Video: vertical (9:16) videos for playback and sharing, assembled in a fixed layout from the lyrics, the AI-generated audio, and the cover image. Apart from the cover image, AI does not draw any new visuals.

2. Underlying AI models

Text generation (readings, lyrics, consultation answers, and similar): DeepSeek V4 Flash (DeepSeek) / Claude Haiku 5.5 (Anthropic) / GLM-5.3 Flash (Zhipu AI). Depending on the feature and time, one or more of these is used.

Song audio and cover image generation: Suno V5, accessed through kie.ai.

BaZiFlow is operated under its own product name. It is not an official app of any model provider above and is not endorsed by them. This page is updated when the models change.

3. Prohibited content

Generating the following six categories of content, or using the service to try to generate them, is prohibited. This policy applies equally to the inputs you submit, such as birth details and questions.

• Pornography / NSFW: sexually explicit or obscene content.
• Violence / gore: excessively cruel or graphic depictions, or content that promotes or glorifies violence.
• Hate speech: discrimination, insults, or incitement of hatred based on race, ethnicity, nationality, religion, gender, sexual orientation, disability, and similar traits.
• Child sexual abuse material (CSAM): any sexual or sexualized content involving minors.
• Deepfakes / impersonation: content that impersonates a real person, brand, or organization, or imitates someone’s voice or likeness without their consent.
• Copyright / trademark infringement: copying or using another party’s lyrics, songs, characters, brand names, or trademarks without permission.

4. Other prohibited conduct

• Content that promotes or facilitates terrorism, extremism, mass violence, or self-harm, or technical instructions for weapons of mass destruction.
• Publishing another person’s birth details or other personal data without their consent, or using them to harass, impersonate, or discriminate.
• Passing AI-generated content off as human-made work when publishing or selling it.
• Inputs designed to bypass safety measures (jailbreaks), automated bulk generation, spam, or deceptive commercial use.
• Knowingly spreading misinformation, or presenting generated content as medical, legal, financial, or other professional advice.

5. How content is checked

This describes what we actually do today, without promising more than that. We do not use a separate content-safety moderation API.

• Input design: the chart song screen has no free-text prompt box. Lyrics are generated from chart data and the song type you pick.
• Generation constraints: the lyrics prompt forbids imagery of death, disaster, and illness, predictions about the future, and ages or dates. Consultation answers are likewise confined to the scope of a reading by their prompt.
• Automated checks: generated lyrics are automatically checked for structure and for leaked terminology, ages, and similar details. This is a quality check, not a determination that content is safe.
• Provider filtering: the song audio provider (Suno) applies its own sensitive-word filter, and lyrics that trigger it are rejected.
• Human review after a report: reported content is reviewed by our team, who apply the actions in Section 7. We do not run round-the-clock manual monitoring.

6. Reporting

If you find generated content that breaks this policy, infringes your own rights, or impersonates you, email support@baziflow.com with the subject “Content Report”. Include the page link and why you think it is a problem, which makes the review faster.

Reporter details are used only to review the report. Malicious or false reports are recorded and may be acted on.

Anything involving child safety is handled first. We aim to give a first response within 3 business days, and serious violations are prioritized. If a review needs longer, we will tell you why.

7. Enforcement

Depending on the content and how serious the violation is, we may take one or more of the following actions.

• Warning: a written notice asking for removal or changes.
• Content removal: confirmed violating content is deleted from the service and its stored files are removed.
• Feature restriction: limits on certain features or on generation.
• Account suspension: a temporary suspension for serious or repeated violations.
• Account termination: for the most serious violations, or when problems continue after a suspension.
• Legal action: highly unlawful content may be reported to the authorities, and we will cooperate with investigations.

8. Appeals

If you disagree with an action, email support@baziflow.com within 30 days of the notice with the subject “Appeal”. We will look at it again and tell you the result.

9. Updates and contact

Material changes are reflected in the last-updated date on this page. Continuing to use the service after a change means you accept the updated policy.

Questions, reports, and appeals all go to support@baziflow.com. For the general terms of the service, see the Terms of Service.