Beta: these terms are being reviewed by counsel and may change.
Legal
Acceptable Use Policy
Last updated October 25, 2026
This policy lists what you can't do with XCap. It applies to everyone in your workspace and to anything our AI features generate at your request.
1. Illegal or harmful content
- No content that is illegal, sexually exploits minors, promotes terrorism or violence, or harasses, threatens or doxxes anyone.
- No hate speech targeting people based on protected characteristics.
- No sexually explicit content.
2. Deception and impersonation
- Don't impersonate real people or organisations.
- Don't present AI-made content as made by a person, or an AI-made person as a real one, to deceive anyone.
- Don't create misleading health, financial or political claims, fake reviews or testimonials, or undisclosed paid endorsements.
- Don't generate election misinformation or content designed to suppress voting.
3. Real people and characters
- No celebrities or well-known public figures — for example politicians and officials, entertainers, musicians, athletes, business leaders, journalists and online creators with a large following — in images, clips, presenters, AI twins, voices or animated photos. This applies even if they, or their estate, give permission.
- Any other real person may appear in what XCap generates only as an AI twin or a voice clone made in XCap with that person's recorded consent.
- No copyrighted or trademarked characters: film, TV, game, comic and cartoon characters, brand mascots, virtual influencers and VTubers. Your own brand's mascot or character is fine.
- AI-made people, such as our stock presenters, are fine as long as they don't copy a real person or a protected character.
- No AI twins, voice clones or animated photos of anyone under 18.
- Scripts and captions may name or discuss real people, for example in news or commentary. They must not fake an endorsement, put words in someone's mouth, or run a giveaway in a celebrity's name.
- We check prompts, photos and what XCap makes for public figures, characters and minors, and refuse what we find. The checks can miss things, so following these rules is up to you.
4. Voice cloning rules
- Only clone your own voice or a voice whose owner gave you explicit, informed consent for that use.
- No cloning of celebrities, public figures, minors or people who have died, even with permission from them or their estate.
- Never use a cloned voice for fraud, scams, robocalls, or to bypass voice-based security.
- Delete a voice when consent is withdrawn.
5. Other people's rights
Only upload or publish material you have the right to use. Don't infringe copyrights, trademarks, or privacy or publicity rights, and don't recreate copyrighted or trademarked characters (see Real people and characters).
6. Spam and outreach
- Follow CAN-SPAM, GDPR, CASL and similar laws. Honour unsubscribes; we suppress unsubscribed addresses automatically.
- Don't upload contact lists you don't have permission to email, or use the Service for bulk unsolicited messages.
- Follow the rules of every platform you post to, including limits on automated posting.
7. Rules from the AI models we use
Two of the AI models behind XCap come with use rules that their licences require us to pass on to you. You must follow them whenever you use XCap, on top of the rest of this policy. Below is a plain summary; if it differs from a licence, the licence's wording applies.
LTX-2 (Lightricks). LTX-2.3 makes our "With audio" clips. It is licensed under the LTX-2 Community License Agreement (read the full text there). Use it only for lawful purposes. You must not use it, or anything it makes, to:
- break any law or regulation;
- exploit or harm minors in any way;
- create or spread false information meant to harm others;
- create or spread personal information that could be used to harm someone;
- publish machine-generated content without clearly saying that it is machine-generated;
- defame, disparage or harass anyone;
- impersonate anyone without their consent, including with deepfakes;
- make fully automated decisions that harm someone's legal rights or create binding obligations;
- discriminate against or harm people because of their social behaviour or their known or predicted personal characteristics;
- exploit the vulnerabilities of a group of people (for example because of their age or a disability) to change their behaviour in a harmful way;
- discriminate against people because of legally protected characteristics;
- give medical advice or interpret medical results;
- produce information for law enforcement, courts, immigration or asylum decisions, such as predicting whether someone will commit a crime;
- create malware or anything meant to harm computer systems;
- promote discrimination or unlawful harm in employment, credit, housing or other essential services;
- help anyone harass, abuse, threaten or bully others;
- support military, warfare, nuclear or weapons work, or anything likely to cause death, injury or serious damage;
- train or fine-tune other AI or machine-learning models;
- get around safety features, content filters or other restrictions;
- build or run a product that competes with or replaces Lightricks' own products.
When you publish a "With audio" clip, label it as machine-generated.
Gemma (Google). Gemma 3 runs inside the same clip engine. Gemma is provided under and subject to the Gemma Terms of Use. Its Prohibited Use Policy says you must not use it, or anything it makes, to:
- produce content that infringes anyone's rights, such as copyright;
- help with illegal activity, including child sexual abuse material, illegal drugs, goods or services, other crimes, violent extremism or terrorism;
- practise law, medicine, accounting or finance without the required licence;
- send spam, or create content for fraud, scams, phishing or malware;
- get around safety filters, or push the model to break these rules;
- promote hatred, harassment, bullying, violence or self-harm; share personal information to harm someone; track or monitor people without their consent; treat people unfairly because of sensitive or protected characteristics; or generate or infer sensitive private information without the rights and consents needed;
- mislead people: claim that AI-made content was made by a person, or present it as original work, to deceive; impersonate a living or dead person without clearly saying so; make misleading claims of expertise, especially about health, finance, government services or law; make automated decisions that affect people's rights or wellbeing (for example in finance, law, employment, healthcare, housing, insurance or social welfare); or make defamatory statements, images or audio;
- generate sexually explicit content (our own rules above forbid it in every case).
8. Protecting the Service
- Don't probe, scan or test our systems for vulnerabilities without written permission, or try to get around rate limits, credit charges, content checks or access controls.
- Don't share accounts, resell access, or scrape the Service.
- Don't upload malware, or use the Service to build a competing AI model from its output.
9. Enforcement and reporting
We may remove content, pause jobs, or suspend accounts that break this policy, and report illegal activity to authorities. To report misuse, including an unauthorised voice clone or an image or video of you or someone you represent, email [email protected].