Frequently asked questions

Navōn (navon.org) is a kosher AI platform for families, schools and frum workplaces. It gives you chat, pictures, video, voice, documents, printable worksheets and study tools from the leading AI models, with every request and every answer checked on our servers against the standards your home or school sets.

There is no hechsher for software, so we don’t claim one. What Navōn offers is what people mean by the phrase: you set the standard, it is enforced on the server for every tool and every device, it refuses rather than guesses when the check can’t run, and an adult can see what was declined.

Adults who want AI without the parts that don’t belong in a frum home, parents who want their children to use it safely, and yeshivos, Bais Yaakovs, seminaries and organizations that want it for students and staff with one policy for everyone.

Navōn uses those same models, so the answers are as capable. The difference is around them: every question is checked before it reaches a model and every answer before it reaches you, against settings a parent or school controls, with a record of what was declined. It also answers within Torah values by default, and draws people modestly.

It can explain sources, translate and unpack a passage, and answer within your community’s practice when you tell it which one. It does not pasken: for a practical question it points you to your rav.

Two checks, both on our servers. The first reads what was asked before any model sees it; the second reads the answer, or looks at the picture, before it’s shown. Anything that doesn’t meet the standard is declined with a kind explanation and, where it makes sense, a better way to ask. Because the checks run on our side, they apply the same way on every device.

The filter doesn’t run on the child’s device and isn’t part of the conversation, so there are no settings to change and no instructions to talk it out of. A clever prompt still has to pass the check on the way in, and the answer is checked again on the way out. No filter is perfect, which is why parents and schools can also see what was asked and what was declined.

Navōn refuses the request rather than letting unchecked content through. It fails closed: you’ll be asked to try again in a few seconds instead of getting an answer nobody checked.

A little. The question is checked before the answer starts, which is a short pause, and the answer is checked as it finishes. Most people notice a second or two; it’s the cost of checking both ways.

A core set of categories is blocked for everyone and can’t be switched off, including explicit and immodest content. Everything above that core is yours to set: how strict, which subjects to keep off, and the tone answers take.

Some passages in Tanach and Gemara touch on subjects the filter normally declines. An adult can turn on Torah-learning mode for themselves, or a parent for a child, which allows those subjects when the question is genuinely about learning a source, and nowhere else.

Every picture is asked for modestly before it’s drawn and checked again after: one that doesn’t meet the standard is withheld rather than shown. People are drawn as observant Jews unless the request says otherwise, dressed the way your community dresses.

If a child writes something that suggests they may be going through something hard, such as distress or self-harm, Navōn answers kindly, points them to a trusted adult, and emails their guardian, or their school’s staff, within moments.

How strict the filter is, subjects to keep off in your own words, allowed hours and a daily time limit, which tools are on (such as pictures or the homework helpers), and Torah-learning mode. You can also see what was asked and what was declined, and get a weekly summary by email.

Yes. Every setting is per child, so a teenager and a younger sibling can each have what suits them.

From the Family Hub in Settings, add or invite each child. A child under 13 needs a parent to set up the account; from 13, they can be linked to you. The filtering follows the child’s account on any phone, tablet or computer.

Navōn is for 13 and over on its own. Younger children can use it through an account a parent sets up and supervises.

No. The filter settings are locked behind a passcode that only the parent knows, and a supervised account can’t loosen what its parent or school has set.

A school can start on its own at navon.org/schools/start: name the school, choose how many student seats, and pay by card. You become the school’s admin and add staff and students from the school console. Prefer to see it first? Ask for a walkthrough on the Schools page.

$8 per student a month, or $80 a year (two months free). Teachers, office staff and parents are free. You can change the number of seats at any time, and it’s prorated. Need an invoice or a purchase order instead of a card? Write to billing@navon.org.

Admins and teachers see what students asked in every tool, Navōn’s replies, and what was refused, with a weekly overview and a list of students worth a look. Students are told plainly in their settings that their school can see this. Staff accounts themselves aren’t monitored.

Yes. Teachers make classes, add their school’s students, set assignments with due dates, and see every student’s work in one gradebook, then return a grade and feedback or ask for a revision. Students write and hand in their work in Navōn.

Yes. A school sets one policy for all its students, or different rules by class, grade or age, including allowed hours and subjects to keep off. The school’s rules follow each student onto any device.

For chat: Claude (Anthropic), GPT (OpenAI), Gemini (Google) and Navōn’s own tuned options, plus Grok (xAI) for adults. For pictures: Nano Banana, GPT Image and Grok Imagine. For video: Veo, Seedance, Grok Imagine and Wan. For voice: ElevenLabs. Every request is filtered the same way whichever model answers it.

No. Leave it on Auto and Navōn picks the right model for each question, and searches the web when the answer depends on recent information. You can always pick one yourself, or compare two side by side.

Answers and writing in chat, pictures, short videos, voiceovers and read-aloud, documents and letters, printable worksheets and signs, small apps you describe in words, and news on a topic from the sources you choose. There’s also an Agent that can search, read sources and check the calendar for you, each step filtered.

Yes. You can type or dictate in Hebrew and Yiddish, answers are laid out right to left, and Read aloud speaks Hebrew as well as English.

Yes. Navōn runs in any modern browser on a phone, tablet or computer at navon.org, and you can also ask it questions on WhatsApp once you link your number in Settings. The filtering travels with your account, not the device.

Yes. Settings has an Import history page that brings in your past conversations from ChatGPT, Claude or Gemini, so you don’t start from nothing.

Not yet. Connecting Gmail, Google Drive and Google Calendar is coming soon, and when it arrives it will be read-only, one item at a time, and only what you choose to bring into a conversation, checked by your filter first. Until then you can attach a file or paste text into any conversation.

Yes. The Navōn API is compatible with the OpenAI SDKs, and every request and answer goes through the same filter as the app. Keys come with a paid plan; see navon.org/developers.

Pro is $19.99 a month for one person. Family is $24.99 a month and includes two seats, with parental controls for each child. Schools pay $8 per student a month or $80 a year, with staff free. Every plan includes filtering on every question and every answer.

Not at the moment. We’re opening in stages, and every account is reviewed before it’s activated, so nothing is charged until yours is open.

Every new account is reviewed by a person before it can be used, which is part of keeping the platform safe for children. Most requests are handled within a day or two, and you get an email as soon as yours is approved.

Yes. You can cancel, change plan or update your card from Billing in Settings, and your plan stays active until the end of the period you’ve paid for.

No. Your conversations don’t train any model, ours or anyone else’s, and we don’t sell your data. To answer you, your question goes to the company whose model writes the reply and to Google for the safety check, and nowhere else.

Only you, unless you’re a supervised account: a parent sees their child’s activity, and a school sees its students’. Supervised accounts are told this in their settings.

Yes. You can turn conversation history off, delete conversations, see and clear what Navōn remembers about you, download a copy of everything, or delete your account for good, all from Settings.

Turn on two-step sign-in in Settings → Privacy & security, so a stolen password alone isn’t enough. You can also sign out every other device from there.