Access and privacy

The link is the first layer

Every lens has a shareable link, and anyone with it can open the lens. The Access section of Lens settings adds more layers on top, and scope decides what the lens can read at all.

What the lens can read

By default a lens reads everything in its workspace. Under Access you can narrow that to only selected spaces; a space scoped into a lens also brings its subspaces. That scope is the wall between your private mind and the slice you are sharing. A lens cannot see, search, or quote anything outside its scope, and there is no way for a visitor to talk it around that.

Sensitive things are always refused

Even inside its scope, a lens will not hand out the things that should not be shared: home and work addresses, phone numbers, passwords, medical or financial details, legal matters, or anything sensitive about other named people. This holds no matter how a visitor phrases the question.

It won’t make things up

A lens answers from your notes, not guesswork. If something is not covered, it says so. It never invents names, numbers, dates, or details about you.

Shape it with instructions

• Custom instructions - short rules the lens follows on every reply (“Always point people to my booking link”, “Never discuss pricing”). They shape style and behavior, but they cannot override the privacy and refusal rules.
• Per-fact guidance - when you capture a fact, you can attach a note about how it should be shared, like “share the city, not the street.” The lens respects that literally.
• Custom refusal - write your own refusal sentence, what the lens says when it cannot answer, so even a “no” sounds like you.

Ask visitors who they are

• Require a name - visitors must enter a name before they can chat. The name shows up in your Inbox so you know who has been talking to your lens.
• Require an email - the address shows up alongside the conversation in your Inbox and in Audience, so a one-off question can turn into a follow-up later. This works on trust: you ask, they type, the chat opens.
• Verified email - flip the second toggle and the email also gets a one-time code. Visitors enter the code before chat starts, which means a returning visitor on a different device picks up their previous conversation instead of starting fresh. One verified email is one persistent identity across sessions.

Password

Add a password and only people you have given it to can get in. Good for a lens you want to share with a specific group, not the open web.

Visitors do not land on a bare password box. The lock screen introduces you first, showing your picture, your name, your greeting, and your vibe words in your lens’s own theme and accent color, and asks for the password underneath. Nothing private is shown: it is the same picture, name, and greeting that already appear in the link preview when you paste the lens into a chat.

Expiry and pausing

Set a lens to expire after a day, a week, a month, or never. When it expires, the link stops working. You can also retire a lens manually at any time from its page, which kills the link immediately, or pause and resume the public link from the top of the Lens page.

A cleaner URL

On paid plans you can claim a slug, a short readable URL instead of the default random one. On higher plans you can connect your own custom domain, verified via DNS, so the lens lives on your brand. The app shows what your plan includes; see Plans and billing.

AI disclosure and the consent gate

Every reply carries a persistent AI badge in the header; see AI disclosure. If your lens speaks for a real person, you can add a stronger one-time consent gate: the first time a visitor lands, they see a blocking message explaining they are talking to an AI version of you and must click Continue before the chat becomes interactive. The moment of consent is timestamped on their session, so you have a record of it.

Enable the gate from the Access section of Lens settings. Leave the message field blank to use the default copy, or write your own (up to 1000 characters). The gate only shows on a visitor’s first visit and does not re-prompt after that. This one-time disclosure is closer to what EU AI Act compliance asks for.

Visitors see a reminder

Below every lens chat, visitors see a one-line note that AI replies come from your knowledge, can be incomplete or wrong, and should be verified before acting on them. It keeps people from treating a single answer as authoritative, especially on lenses about medical, legal, financial, or other high-stakes topics.

Try this

• Sharing publicly? A slug plus a photo makes the link feel legit. Sharing privately? A password plus a name gate keeps it tight.
• Before publishing, open the lens link yourself and try to get it to leak something. It is the fastest way to trust your scope.