The short version
Not a substitute for the rest of this page, but an accurate summary of it:
- Naven is free. There is no paid tier, no card, and no advertising.
- You need to be 13, or have a parent or guardian approve your account.
- Every message, picture and voice note is checked before it sends,
against seven published categories of conduct, plus slurs and strong language. Disagreements and insults are not among them.
- Because Naven checks messages, it can read them in transit. It is
not end-to-end encrypted, and if that is your first requirement you
should use Signal instead. We would rather say so here than have you find out later.
- Your messages are yours. Naven does not sell them, train on them, or show them to
advertisers, because there are no advertisers.
- Naven is a free product built by one person. It can break, and it comes with no
guarantee that it will be available.
1 · Who may use Naven
You may use Naven if you are 13 or older. An account that says it
belongs to someone under 13 may use Naven, and Naven applies its strictest screening to
it. If that account repeatedly sends something the safety screen refuses — swearing
counts here, where it would not for an adult — messaging is suspended until a parent or
guardian confirms the account by clicking a link Naven emails to their own inbox. That is
a real gate, not a checkbox: it cannot be completed by the child, and for an under-13
account it is the only way through — the self-verification route offered to adults is
not available.
The age you give at sign-up sets your protection level. Naven stores only the band
(under 13, teen, adult), never your birth year. If Naven has reason to doubt the age on an
account — see section 6 — it may pause sending until the age is
confirmed.
You may not use Naven if a previous account of yours was removed for breaking these terms.
2 · Your account
You can sign in with Google, with an email address, or as a guest. A guest session
lives only in that browser: clearing your browsing data destroys it, and
it cannot be recovered by anyone, including us. If an account matters to you, attach an
email address to it.
Keep your sign-in to yourself. You are responsible for what happens on your account, and
if you think someone else has access to it, sign out of other sessions and contact us.
Naven has no public member directory. Nobody can browse a list of members, and your email
address is never shown to another member — the only thing a lookup can reveal is your
display name, and only if you turned that on.
3 · What you agree not to do
The prohibited conduct is the same ten categories Naven screens for. They are written
out in full because a policy you cannot read is a policy you cannot follow:
- Spam and scams — phishing, fake giveaways, posing as a bank, a
service or a relative to extract money, codes or credentials, and sending the same
unsolicited message to people who did not ask for it.
- Malware and malicious links — links or files intended to compromise
somebody's device or account.
- Harassment and bullying — threatening violence against a person,
telling someone to end their life, or sustained degradation aimed at a person.
- Hate speech — slurs, and attacks calling for harm against or
dehumanising a protected group.
- Credible threats of violence — expressing a serious intention to
harm another person.
- Sexual exploitation or abuse — requesting sexual images, sextortion,
grooming, and anything sexual involving a minor.
- Illegal transactions — arranging a drug sale, stolen data, forged
documents, untraceable firearms, or violence for hire; and recruiting for a violent
extremist group.
- Doxxing and privacy violations — publishing another person's home
address, phone number, workplace, school or identification without their consent.
- Encouraging self-harm — encouraging or instructing another person to
hurt themselves. Describing your own distress is not prohibited and is never
blocked; see the note below.
- Graphic violence — real footage of a person being killed or maimed,
and imagery used to threaten or shock.
Screened in addition to the list above: slurs, and strong language.
Both are refused for every member, at every protection level, with the reason shown to
whoever wrote the message.
Deliberately not screened: insults, disagreements, political opinions,
dark humour, sexual conversation between adults, talking about drugs, describing a violent
film, and someone saying they feel awful. Naven delivers those. A messaging app that
refuses to carry "I hate you" or "I want to die" is not protecting anybody — the second
one is a person reaching for a friend, and stopping it does harm. When Naven recognises
that message it delivers it and offers a helpline alongside it, and nothing about it is
recorded against the account.
Display names are held to the same ten categories and to three additional
rules: a name may not pose as Naven staff or an administrator, may not be a link, email
address or phone number, and may not be built from invisible or direction-reversing
characters. A name that fails is replaced with a neutral placeholder wherever Naven
renders it.
You also agree not to:
- Use Naven to break the law, or to help someone else break it.
- Automate, scrape, or hammer the service in a way that degrades it for others.
- Attempt to defeat the screening, the age gate, or another member's privacy settings.
- Upload malware, or files you do not have the right to share.
- Resell or rebrand Naven as your own service.
4 · Screening before a message sends
What this section describes is an attempt, not a guarantee. Screening
misses things, including obvious things, and it is much weaker in languages other than
English. Section 9 sets out the limits in detail and it is the section that governs.
Naven checks a message before it is delivered rather than after it is reported. So are
pictures and voice notes: a picture is examined by an image model, and a voice note is
transcribed so the same rules can be applied to speech as to typing. That transcript is used for
the decision and then discarded — it is never stored, never shown to anyone, and
never attached to the message.
If something is blocked you are told immediately and told why, and the message is not
delivered. Nothing is queued for review behind your back.
The cost of this, stated plainly. Screening a message and being unable to
read it cannot both be true. Naven chose screening, so Naven's servers can read message
content in transit. Messages are encrypted in transit and at rest, but they are
not end-to-end encrypted. If end-to-end encryption is your first
requirement, Signal is the better product and we recommend it without reservation.
5 · Enforcement
A blocked message is recorded. Three blocks within a fortnight pauses sending on the
account until the age on it is confirmed — by a guardian for an account under 13, or by a
code sent to the account's own address otherwise. Confirming clears the record; it is a
gate, not a punishment that accumulates forever.
For conduct at the severe end of the seven — anything sexual involving a minor, credible
threats — Naven may suspend or remove an account immediately and without notice, and may
report it to the appropriate authorities.
If you think a block was wrong, email us. The screening is a mix of pattern matching and
machine classification, and both are capable of being wrong.
6 · Your content
What you write stays yours. You keep every right you had in it. You give
Naven only the narrow permission it needs to operate: to store your content, transmit it
to the people you sent it to, screen it under section 4, and display it back to you. That
permission ends when the content is deleted.
Naven does not:
- Sell your messages, or any data derived from them.
- Use your messages to train models. Screening classifies a message and discards it;
it does not become training data.
- Show your content to advertisers. There are none.
Content posted to a public channel is visible to everyone who can see that channel, which
may be anyone with the link. Treat it as public, because it is.
8 · Naven AI and the newsroom
Naven includes an assistant and an automated newsroom. Both are machine-generated and
both can be wrong.
- Not professional advice. Nothing the assistant says is medical,
legal or financial advice. For anything that matters, ask a professional.
- Check anything that matters. The assistant cites sources where it
has them. Follow them.
- The newsroom has a floor. A briefing publishes only when at least
five distinct established publishers have reported the story, and every publisher is
listed underneath it. That is a real editorial rule, not a slogan — but it is a rule
about corroboration, not about truth.
Your conversations with the assistant are private to your account. They are not shared
with other members and are not used for training.
9 · What Naven does not promise
Read this section before you rely on Naven for anything that matters.
Naven is new, it is free, and it is written and run by one person. It makes mistakes,
including serious ones, and the paragraphs below are not boilerplate — they describe how
this software actually behaves.
Naven does not promise security. Not "military-grade", not
"bank-level", not any level. Messages are sent over HTTPS and stored on infrastructure
run by Google and Cloudflare, and that is the whole of it. Naven is
not end-to-end encrypted: Naven's servers can read message content, and
so, in principle, can anyone who obtains access to them. Do not send anything through
Naven that would harm you if it were read by someone else, published, or lost.
Naven does not promise moderation. Screening runs before a message
sends, and it is best-effort software written by one person, not a guarantee and not a
service being sold to you. It misses things. It misses obvious things. Specifically, and
without this being a complete list:
- Other languages are much weaker. The rule-based part of the screen is
written in English. Anything in another language is left to a language model alone,
which is far less reliable — so harmful content in Spanish, French, Persian, Russian,
Chinese or any other language is substantially more likely to get through
than the same thing in English. Depending on the message, it may not be
screened in any meaningful way at all.
- Meaning across several messages is often missed. The check at send
time looks at one message. Something harmful assembled across four is frequently not
caught, and the later review that looks for this runs once a day, catches only some
of it, and can be wrong in both directions.
- Slang, coded language and new abuse are missed. The screen knows what
it was written to know. It does not know what was invented last week.
- Pictures, voice notes and video are weaker still. A picture is judged
by a model. A voice note is transcribed first, and transcription is unreliable —
accents, background noise, quiet speech and any language other than English all
degrade it, and a bad transcript is screened instead of what was said.
- Calls are not screened at all. Live audio and video pass through
untouched. Nothing looks at them, during or after.
- It blocks things it should not. Ordinary messages are refused, and
that will happen to you.
Naven is not a child-safety product and must not be relied on as one.
Parental controls, age bands and screening are conveniences that can fail, be
circumvented, or simply not work. They are not supervision, they do not replace it, and
no combination of settings makes Naven safe for a child to use unsupervised. A
determined adult can reach a child on Naven, as on any platform where people can message
each other.
Age checks establish very little. Age bands come from what somebody
typed at sign-up. The provider check confirms only what that provider says, which Naven
cannot verify and does not audit. Anybody may be any age.
And the ordinary ones:
- No uptime guarantee. Naven can be down, slow, or partly broken. There
is no service level agreement and no compensation for downtime.
- No guarantee of delivery. Messages usually arrive immediately. They
may not arrive at all, and Naven may not tell you.
- No guarantee of retention. Messages, pictures, voice notes and
accounts can be lost. There is no backup you can ask for and no recovery process.
Keep your own copy of anything you need.
- Features can change or be withdrawn. Including ones you rely on, and
without notice.
- Third-party services can fail. Calling, email, storage, the assistant
and the newsroom all depend on outside providers, and their failures are not within
Naven's control.
- Naven may remove content after it was delivered. A later review can
retract a message from every device that received it.
To the fullest extent the law allows, Naven is provided "as is" and "as available",
without warranties or conditions of any kind, express or implied — including any implied
warranty of merchantability, fitness for a particular purpose, non-infringement,
accuracy, or that the service will be uninterrupted, secure, or free of harmful content.
No advice or information obtained from Naven or its developer creates any warranty not
stated here. Neither Naven nor its developer is liable for indirect, incidental, special,
consequential or exemplary losses, nor for loss of data, loss of profits, or harm arising
from content sent or received through the service or from screening that failed to
prevent it. Nothing here limits liability that
cannot lawfully be limited — including for death or personal injury caused by negligence,
or for fraud. Some jurisdictions do not allow these exclusions, in which case they apply
to you only as far as that jurisdiction permits.
10 · Ending it
You can leave at any time. Email us to have your account and its content
deleted, and see the Privacy Policy for exactly what is removed and
what necessarily remains (a message you sent to someone else stays in their conversation —
Naven cannot reach into another person's inbox and unsend your side of a chat they took
part in).
Naven may suspend or end your access if you break these terms, if keeping your account
would put other members at risk, or if Naven shuts down. If Naven shuts down, we will give
as much notice as is practical and a way to export your data.
How a suspension actually works, written out because you are entitled to know
before it happens. An ordinary first block is a warning. A second within thirty
days is a one-week suspension. The most serious categories — grooming, sexual
exploitation, threats and blackmail — start at two to four weeks rather than at a
warning, because a warning is not a proportionate answer to them. Each further block
while suspended extends from the existing end date and moves up the scale, to a maximum
of five weeks at a time. A clean month clears accumulated warnings; it does not reset a
suspension level you have already reached.
An account that has completed age verification and then posts blocked content again
enters that scale immediately rather than starting with a warning, because the check has
already been explained to that member once.
An administrator can also block an account by hand, which overrides the
automatic scale in both directions. When that happens Naven shows you the reason, any
message the administrator wrote to you, and the date it ends. A block with an end date
lifts by itself on that date — you do not need to ask, appeal, or sign in again for it to
clear. Signing out or creating a second account does not lift a block, and a second
account made to get around one may itself be closed. If you think a block is wrong, reply
to the email Naven sent you or write to
help@navenapp.com.
Display names. Naven screens display names to the same standard as
messages and to three additional rules — a name may not impersonate Naven staff, may not
be a link, email address or phone number, and may not use hidden or
direction-reversing characters. An administrator may replace a name that breaks these
rules, and a replaced name cannot be changed back by the account holder.
11 · Changes to these terms
These terms will change as Naven does. The date at the top always reflects the current
version. For a change that materially reduces your rights or expands what you agree to,
Naven will tell members with an email address before it takes effect. Continuing to use
Naven after a change means you accept it; if you do not, stop using it and ask us to
delete your account.
Naven