Acceptable Use Policy

This Acceptable Use Policy applies to every user of Bento and to prompts, instructions, reference files, images, video, audio, text, metadata, and other content submitted to, generated by, edited with, or distributed through Bento.

You are responsible for your content and must have all rights and consents needed to use it. AI generation does not remove that responsibility. Do not attempt to evade a provider’s safety system or Bento’s safety controls.

You may not use Bento to generate, edit, request, upload as a reference, or distribute any of the following six categories:

• Sexual or NSFW content: pornography, explicit nudity, sexual services, fetish content, or sexualized depictions of a person without consent.
• Violence or graphic gore: graphic injury, torture, mutilation, executions, glorification of extreme violence, or content intended to shock through realistic gore.
• Hate speech: content that attacks, dehumanizes, threatens, or promotes discrimination or violence against people based on protected characteristics.
• Child-unsafe content: child sexual abuse material (CSAM), grooming, sexualized or exploitative depictions of minors, or content that facilitates harm to a child.
• Deepfakes or impersonation: deceptive or non-consensual use of a real person’s face, voice, likeness, or identity, including impersonation used for fraud, harassment, sexual content, or political deception.
• Copyright or trademark infringement: content that unlawfully copies protected works, characters, logos, brands, trade dress, or other intellectual property, or falsely suggests endorsement.

Bento has zero tolerance for CSAM or any sexual content involving minors; content that promotes terrorism, violent extremism, genocide, or mass violence; instructions for weapons of mass destruction; and non-consensual sexual deepfakes. We may preserve relevant records and report suspected illegal activity to the appropriate provider or authority when required or permitted by law.

You also may not use Bento for fraud, phishing, malware, cyberattacks, doxxing, stalking, unlawful surveillance, threats, harassment, self-harm encouragement, regulated professional deception, misinformation intended to cause material harm, spam, or attempts to jailbreak or bypass safety controls.

Bento-hosted generation is introduced through Early Access. A model is not eligible for general hosted availability until its generation path includes safety controls for both user input and generated output.

For models made available through Bento-hosted generation, the moderation process includes automated screening of prompts and relevant reference information before submission, provider safety controls and output screening before a result is made available, and manual review by the Bento operator when a request or result is flagged or reported. A request may be blocked, delayed, or withheld when screening is unavailable, inconclusive, or indicates a policy risk.

Bento periodically reviews blocked requests, abuse patterns, user reports, and the coverage of its safety rules. Because Bento is local-first and does not currently operate a public content gallery, local project files are not uploaded solely for moderation. Requests made directly with your own API key may also be governed by your chosen provider’s safety policy.

Report suspected violations to founders@bentostudio.net with the subject “Content Safety Report.” This mailbox is monitored by the Bento operator.

Include the relevant task or transaction ID, model name, approximate date and time, a description of the issue, and a URL or screenshot when safe and lawful. Do not download, attach, forward, or redistribute suspected CSAM; instead provide only the task ID, location, and circumstances so it can be handled safely.

We keep a reporter’s identity confidential where reasonably possible, review reports neutrally, and may restrict malicious or knowingly false reporting.

• Critical safety reports, including child safety, terrorism, credible threats, and non-consensual sexual deepfakes, receive priority review with a target initial response within 24 hours.
• High-risk reports, including graphic violence, targeted hate, fraud, or repeated impersonation, have a target initial response within 2 business days.
• Other policy and intellectual-property reports have a target initial response within 5 business days.

Complex cases may require additional time, identity or rights verification, consultation with a provider, or preservation of evidence. We will communicate material delays where reasonably possible.

Depending on severity, context, history, and legal obligations, Bento may block a request or result; issue a warning; remove Bento-controlled copies of prohibited content; restrict generation features, models, rate limits, or Credits; temporarily suspend an account; permanently terminate access; preserve evidence; or report suspected illegal conduct to a provider or competent authority.

Critical violations may result in immediate permanent suspension without a prior warning. Repeated or deliberate attempts to evade safety systems may receive a more serious response. Paid balances and refunds will be handled under the Terms of Service and applicable law; promotional or fraudulently obtained Credits may be revoked.

If you believe an enforcement decision was wrong, email founders@bentostudio.net within 14 days with the account email, decision notice, and reason for appeal. We aim to complete an appeal review within 10 business days. Where practicable, the appeal will be reviewed separately from the initial decision.

We may update this policy as models, risks, and legal requirements change. Material changes will be announced through email, an in-app notice, or a prominent website notice. Questions about this policy can be sent to founders@bentostudio.net.