Terms of Service

Acceptable Use Policy — AI voice and content

This policy forms part of AutoBracketMaker's Terms of Service. It applies to tournament and player names, venue names, announcement text, synthesized audio, and content submitted, shared or played using the service. Use announcements for legitimate tournament operations, obtain necessary permissions and tell listeners that cloud voices are AI-generated. You remain responsible for the content you provide and how you use the output.

AI provider and specific model

Cloud text-to-speech uses OpenAI's gpt-4o-mini-tts model through the OpenAI Audio Speech API. Announcement text, including player and venue names, is sent to OpenAI for synthesis. Audio is AI-generated, not a recording of a human announcer. We offer preset voices, not custom voice cloning or impersonation. The device-voice option uses your browser or operating system's speech service and does not call this OpenAI model through our server. AutoBracketMaker does not persist generated audio.

Six prohibited content categories

• Pornography / NSFW: do not generate or broadcast sexually explicit, pornographic or sexual exploitation content, including non-consensual sexual content.
• Violence / gore: no graphic descriptions of injury, violent threats, incitement to violence, terrorism, or instructions facilitating physical harm.
• Hate speech: no dehumanization, abuse or incitement to discrimination or violence targeting people based on protected characteristics such as race, religion, nationality, gender, sexual orientation or disability.
• Child-unsafe content / CSAM: no child sexual abuse or exploitation material, sexualization of minors, grooming, or promotion of harm to children, including synthetic or fictional depictions. There are no exceptions for child sexual exploitation.
• Deepfakes / impersonation: no deceptive imitation of a real person, brand or organization, unauthorized voice imitation, fraud, or claims that synthesized speech is an authentic recording or endorsement.
• Copyright / trademark infringement: do not submit, reproduce or distribute protected material without the necessary rights, or use names and trademarks to falsely imply affiliation or authorization.

Do not disguise prohibited material as a player or venue name, split it across requests, evade safeguards, or use announcements to harass, deceive or disclose private information. These restrictions apply to both cloud and device playback.

Content controls and review

Public cloud previews accept only predefined samples. Custom cloud announcements require sign-in; the server validates request type, text length, language and preset voice, and applies request limits. These controls restrict access and request shape; they are not semantic screening for all prohibited content.

Content reports are reviewed manually by the site operator using the reported page, context and available records. We do not currently run a separate automated semantic moderation API or review every generated audio output before playback. We cannot guarantee detection of every violation. Because generated audio is not archived by us, a report should identify the relevant tournament and incident time.

Handling violations

Depending on severity and recurrence, we may require correction, issue warnings, remove content under our control, or discontinue access to the service under these Terms. Child exploitation, credible threats and other serious harm receive priority and may require immediate protective action. We cooperate with lawful authority requests and report suspected illegal material where legally required. We cannot recall audio already played or copies outside our control.

Report unsafe content or infringement

Email our existing support address below with the subject 'Content report'. Include the tournament/page URL or ID, incident time, relevant category, a concise description and a reply address. For copyright or trademark complaints, identify the rights holder and the affected work or mark. The operator reviews the information and may request clarification before deciding what action is appropriate.

Do not send CSAM, graphic illegal material, passwords or unnecessary personal data. Provide a location and description instead. Report information is used for investigation and shared only as needed to handle the issue or meet legal obligations. This mailbox is not a 24-hour emergency service; contact local emergency services if someone faces imminent danger.

Appeals and policy updates

To dispute an action, reply to the notice or email the same address with the subject 'Content appeal', the affected page/account, the decision and your supporting explanation. The operator will reconsider the available evidence and communicate the outcome where possible. Policy changes are published here with an updated date; this policy does not promise a dedicated review team or fixed response time.