Apps by Nikki is a two-person studio. We build our own apps, and we build them for other people too.
OUR APP iPhone
Don't rely on one AI. Ask the panel.
One AI can be fluent, persuasive, and wrong. Asking several by hand is tedious, and lining up four chat windows just hands the comparing back to you.
The Panel does the comparing. Each AI answers your question on its own, then you get one verdict: the points they share, where they split, the minority view, and every original answer in full.
CLIENT WORK Website
An online gallery for ELLE and Esquire's Hollywood exhibition.
Built for Osmos.
Hearst Galleries put on a joint exhibition at the Hearst Tower in New York: Hollywood photography from the ELLE and Esquire archives.
The site takes it online. Photographs drift in around the title, then you pick a gallery, browse its wall, and open any picture for its issue, photographer and story.
Nikki has an eye for apps that look great and feel even better, and always wanted a studio of her own. Romeo builds things. The name was her idea, and he's fine with it.
We design and build mobile apps and websites. Some are our own ideas, some are for clients. Either way they get the same fuss over how they look and how they feel.
Between our own apps, we take on freelance and consulting work. If you need one of these built properly, get in touch.
iPhone and Android apps, from the first screen to the app stores.
Websites and web apps, from the first sketch to the day they go live.
Have a project in mind, or a question about our apps? Drop us a line.
The Panel
Effective date: 7 October 2026
Last updated: October 2026
This Usage Policy is part of our Terms of Use (see Terms §5). It covers everything you send The Panel: your questions, any context you add, and any AI answer you paste in for a Second Opinion.
You can't use it to:
(a) Child safety: seek, create or share sexual content involving anyone under 18 in any form, including writing, images, video or live-streaming, or anything that sexualises a child.
(b) Violence and weapons: plan or encourage violence against people, including terrorism, or get help making weapons that can cause mass harm.
(c) Crime and fraud: get help committing a crime or fraud, making or buying illegal drugs or weapons, creating malware, or getting into systems you don't have permission to access.
(d) Threats and hate: threaten people, or encourage violence against people because of who they are.
(e) Getting around our protections: evade our safety check, limits or security, or open a new account after we've closed yours.
(f) Sexual content and self-harm instructions: generate pornography or other sexually explicit material, or content that encourages, or gives instructions for, suicide, self-harm or disordered eating. You can still ask about these subjects for help, health or study (see section 2).
Australian online safety law calls the most serious of this material "class 1A" (child sexual exploitation material, pro-terror material, and extreme crime and violence material) and "class 1B" (crime and violence material, and drug-related material). None of it is allowed on The Panel. If The Panel is used to seek or generate child sexual exploitation or pro-terror material, we may restrict or pause your use for a set period, close your account, and remove the material, as sections 3 to 7 describe.
These rules also reflect the acceptable-use requirements the AI providers who answer your questions place on us. You can ask about difficult subjects for study, news, history, health, law, safety or your work. Our checks are automated and can be wrong, so we warn you before we close an account, and a person reviews any decision you appeal.
You can always ask about suicide or self-harm. On its own, that never counts against you. If our check sees signs that you or someone you're worried about may be at risk, we show you support resources. If the Panelists decline to answer such a question, that uses a Panel Check like any other run where the Panelists don't answer (section 4).
Once you have allowed AI provider sharing, an automated safety check, run for us by OpenAI, reviews your question before the Panelists see it, and reviews each Panelist's answer and the Panel Verdict before we save or show them. Panelists can also decline to answer, and an AI provider's own filter can block a question.
If two or more of the four Panelists decline, are blocked, or have their answers withheld, there's no Panel Verdict and we show none of the answers. This uses a Panel Check, the same as any other run that completes, including when the Panelists declined a question about suicide or self-harm, and including when the step that produces the Panel Verdict was itself blocked or withheld. A Panel Check is not used only when:
A question counts against your account only when both of these happen: (a) our check is highly confident that what you wrote yourself, not text you pasted in, falls under section 1(b), (c) or (d); and (b) at least two Panelists decline it themselves or have their answers withheld by our check. If one already counted in the last 24 hours, another doesn't. Each expires after 90 days.
We keep the words you wrote in those questions, and only those, while they count and while an appeal about them is open, even if History is off, so you and anyone reviewing your appeal can see them. We delete them if you delete your account.
We stop every question that may involve sexual content about someone under 18 (section 3).
We don't keep the text of stopped questions or answers, only the record and the account information described in our Privacy Policy. We report to authorities where the law requires it, including the Philippine Department of Justice and the US National Center for Missing & Exploited Children (see our Privacy Policy, §7).
We close an account only as sections 5 and 6 describe, for a serious attack on our systems or security, or where the law requires it. Our automated checks flag an account; a person makes the closure decision. Section 8 explains how to challenge one. After closure you can still sign in to export your data, delete your account, and manage or cancel your subscription. You can't start Panel Checks or open another account with the same Apple ID. We keep keyed, non-reversible fingerprints of your sign-in identity and your App Store purchases, so a closure stays in place if you delete your account: for 3 years, or 6 years after a child-safety closure. Our Terms (section 11) explain what happens to your Panel Checks and subscription.
When a question counts against you, or we close your account, the app tells you what happened, which rule applied, when, and whether it was automatic, along with a reference code.
To appeal, email support@appsbynikki.com and quote that reference code, within 90 days of the decision, telling us anything you want us to know. A person reviews every appeal, and we aim to decide within 14 days; we'll email you the outcome. While an appeal about a closure is open, the closure stays in place.
If we got it wrong, we reopen your account, remove what counted against you, and restore any Panel Checks that mistake used. Deleting your account ends any open appeal. You can also complain to a consumer regulator or go to court.
An AI provider can refuse questions from your account on its own, and we can't reverse that. If that stops us providing The Panel to you, we'll let you know. That isn't a finding that you broke these rules.
You can report any answer or Panel Verdict with Report a response in the result's menu, or by emailing support@appsbynikki.com yourself. Report a response starts that email with the run's reference and the app version filled in, and a line asking what was wrong, none of what you wrote, so you can say it in your own words. A person reviews reports, normally within 14 days, and removes content that breaks these rules; if you report sexual content involving a minor, we remove it at once. Reporting never counts against you.
We'll tell you in the app at least 30 days before a material change to this policy takes effect, and why. If you don't accept it, you can refuse it and cancel at no cost, as our Terms of Use (§16) explain.
Australia's eSafety Commissioner is the national independent regulator for online safety. eSafety can investigate complaints about seriously harmful online material, including child sexual exploitation material and pro-terror material, and about whether online services meet Australia's online safety codes and standards.
