Flisk Privacy Policy

Last updated: 20 November 2026.

This says what Flisk keeps about you, what it's for, who else sees it, how long it stays and how to remove it. It covers flisk.tv, the Flisk apps, the flisk command line and Flisk's connections for agents. Once, a game made with Flisk, has its own page at its own address.

Who we are. Flisk is run by MakeMoreDigital Limited, the data controller, a company registered in England and Wales (number 07738113) at 20-22 Wenlock Road, London N1 7GU. Write to us at hello@flisk.tv about anything here.

The short version

What we keep, and why

Your account. We keep your sign-in email and ask you to choose a public handle. A nickname is fine. People who can find you on Flisk can search for your handle and any profile name you add. Your profile name and one-line bio are optional. We also keep who you follow, who follows you, your settings, a hash of your password if you have one, your passkeys, and hashes of your access tokens (each with its first few characters and when it was last used, so you can tell them apart). Sign-in codes are kept as hashes and work for ten minutes. We keep this to run your account.

Your age. Where your phone can say how old you are (Apple's Declared Age Range on an iPhone, Google Play's age signals on Android), Flisk's app asks it, and you choose whether to share. It only learns which range you're in (under 13, 13 to 17, or 18 or over), never your date of birth. The app keeps that range on your phone. If you're under 13, it doesn't let you sign in. If you're 13 to 17, it tells us you're under 18. If you're 18 or over and haven't told us yet, it tells us that. If your phone doesn't say, email sign-in asks you to confirm you're 13 or over; we don't store that answer, or ever ask your date of birth. The first time you open Chat, or when a friend invite is waiting on it, we ask whether you're under 18 or 18 or over, unless your phone has already said. We keep the answer and when you gave it. Unless your phone said, it's your own word and we don't check it. It decides who can find and message you (Children).

Your screens. For each screen you add: its name, what it is (its platform, size and what it can draw), its browser or app description, which room it's in, whether it's online and when it was last seen. To let people send to a screen near them, we keep a keyed hash of the internet address it last connected from, never the address. Only the hash's match is used.

What you make. Rooms and scenes with every version of them, what a room's script keeps (its state and storage), what it last fetched, your library of pictures, models, videos and sounds, your streams, your playlists, what you've liked and saved, and the last 100 things you visited. Secrets you give a room are sealed with a key we hold, and only used for that room's fetches. We also keep a record of changes made to rooms and projects, with who made them.

Your own records. What you or an app put in your store, such as workouts or meals. Health records are kept only with your explicit consent, and sealed with a key of your own that we hold. Withdraw consent and no more are written.

Conversations with Flisk's agent. What you say, what it answers and what its tools did, kept so that you can come back to them, and which room each is about. They go to our AI provider (below).

What you add to what you say to the agent. Photos, videos, sounds you record or choose, and files (3D models, music, text, PDFs), when you add them with +, so the agent can make things from them. A photo is kept as a copy without its hidden details (where it was taken, the phone that took it). Only you and the agent in your own conversations can see them. If you ask, the agent keeps one in your library, like anything else you add there; the rest are removed after 30 days.

Messages to other people. The words and any room you share, who sent them and when, whether you've read them, and what you've deleted for yourself. Each person's settings, blocks and message requests. They're kept so that the conversation is there for both of you.

Reports. If someone reports you or a message of yours, or you report someone, we keep who reported whom, the reason and any note, a copy of the reported words, and a label for the reported person (their display name and email). That's so that our staff can judge it, even after the message is deleted. Staff can then suspend someone from messaging, or hide a message from everyone. The person reported is never told who reported them.

Your contacts, only if you tap Check my contacts. The app turns your contacts' email addresses into SHA-256 hashes on your phone and sends those. We compare them with people who've turned on Let people with my email in their contacts find me (off by default, and limited further for under-18s) and answer with any matches. We don't keep what you sent, or that you asked. Your contacts' names, numbers and addresses never leave your phone.

Paying. When you buy credit: the store's record of the purchase, what you were given, and your credit balance and each run's cost and usage. Apple, or on Android Google Play, takes the payment; we never see your card. If you give us your own Anthropic key it's sealed, used only for your runs, and never shown again.

Your phone's push token, to send you notices, and the last time one was sent or failed.

Which days you use Flisk. One row for each person, platform and day (the iPhone app, Android, the web, the command line, MCP, the API, or a kind of screen). No addresses, pages, or what you did. We count sign-ups, active people, returning people and costs from this and from the records above. Only our staff can see it. Our reason is a legitimate interest in knowing how Flisk is used, so we don't ask for consent, and it needs no cookie.

Server logs. Google Cloud records requests to our servers, including your internet address and the address asked for. They're kept for 30 days, in Google Cloud Logging. We use them to run and secure the service.

Our legal grounds:

Your phone's camera, microphone and photos

Who else sees it

Other people on Flisk see your name, handle, bio and follower counts where they can find you. Your rooms are shown to nobody unless you choose. When you share a room in a conversation, they see that version of it, run with none of your data, state or secrets: if it can't work without them, they see a picture of it instead, drawn the same way. They can see what's in the scene itself, so don't put private words in a room you share.

Our AI provider: Anthropic. What you say to the agent, the conversation so far, the room it's about and what its tools return go to Anthropic (in the US) so that its model can answer, on your own key if you've given one. Also: pictures of your screens when the agent looks at them; photos you add to what you say (a smaller copy, and a still from a video you add), and the text of text files and PDFs you add (sounds, videos and other files themselves aren't sent); a room's definition when it asks for more than the sandbox, for a review; and a stream's items if that stream has a "write" step you've set up. Under Anthropic's commercial terms it keeps these for no more than 30 days, unless they're flagged for breaking its usage policy, and doesn't train its models on them. Your messages to other people are never sent. A test keeps it so: no tool of the agent's or MCP's can read them.

Connected agents. If you connect your own agent with a token, it can do what the token allows, and its maker's terms apply. It can't read your messages to other people.

Pictures, videos and 3D models: My Art. When you ask Flisk's agent to make a picture, a video, a 3D model, a cut-out character or a set, the description it writes from what you asked, and any picture of yours it starts from, go to My Art (my-art.space), which passes them to fal.ai (in the US) and the maker of the model chosen for that request (for example Black Forest Labs, ByteDance or Kuaishou). The result comes back to your library. My Art keeps the request and its result in Flisk's account there; we can delete them. Your name, email and messages are never sent: only the description and the pictures the request needs.

Other model providers. Flisk can also be set up to make pictures, sound, music and video with OpenAI, Google or ElevenLabs. None of them is switched on today, and we'll update this page before one is. When the agent searches for media, your search goes to Openverse, Wikimedia Commons, the Internet Archive or NASA.

Our voice server. Words for characters' lines and speech, and descriptions of new voices, go to a server of ours with a graphics card (gpu.makemoredigital.com). It keeps the words only for that request, logs only sizes and timings, and keeps the voices themselves. It's hosted by Hetzner in Helsinki, Finland. Without it, lines go to Google's text-to-speech.

Our motion and photo services. Two more services on the same server of ours. When you ask the agent to act out a movement from a video (yours, or one made with My Art), that video goes there to find the people's poses in it; the moves come back as an animation in your library, and the video is deleted from it as soon as we have them (at most a day, if something goes wrong). When the agent measures a photo of yours to set a scene up from it (where the camera was, the ground, the things and people in it), a fresh copy of the photo, without its location or camera details, goes there and nothing of it is kept on that server. What's measured, people's poses among it, is kept with your library, so the same photo isn't paid for twice. Both log only sizes and timings, and neither can reach the internet.

Our render service. Draws a picture of a scene without a screen, for Studio and previews: a private Google Cloud service that only our servers can call. It keeps nothing but a note of the size and how long it took.

Push notices: Apple and Google. A message's push carries the sender's name and up to 120 characters of the message, through Apple (iPhones) and Google's Firebase Cloud Messaging (Android), to your phone. A message request's push says only who it's from, never its words. A shared thing's says its title. Turn notifications off in your phone's settings to stop them.

Email: Postmark. Sign-in codes, and word to our staff of a new report, which never includes what was said or who reported it.

Usage statistics: Google, only if you say yes.

How often a shared Flisk is seen. Its share page and its embeds on other sites tell us when they're looked at and played, and on which site (the site's name alone). We keep only a count a day for each Flisk, which its owner sees. No cookie is set and nothing about who you are is kept: to count you once in half an hour we keep a scrambled key made from your address in memory for that half hour, then it's forgotten.

Google Cloud runs our servers, database and file storage: in London (europe-west2), files included.

Our staff can see what the service needs them to: reports, accounts, and the admin pages. Messages are not in the admin; staff see only what's been reported. We may share information if the law requires it, or to protect someone from serious harm. We report child sexual abuse and exploitation to the National Crime Agency, and to NCMEC where people in the US are involved, as our Child Safety Standards say.

Moving outside the UK. Anthropic, Apple, Google, Postmark and fal.ai handle some of it in the US, and our voice server is in Finland. Finland is covered by the UK's adequacy regulations for the European Economic Area. For the US, we rely on the UK Extension to the EU-US Data Privacy Framework where the provider is certified under it, and otherwise on the UK's International Data Transfer Addendum in the provider's data processing terms.

Cookies and what's kept in your browser

Signed-in pages use a session cookie and a CSRF cookie, which are needed for signing in. Your cookie choice is kept in your browser's storage for 12 months. Google's analytics cookies (_ga, _ga_...) are set only if you accept them. There are no advertising cookies.

How long

WhatHow long
Account, rooms, library, records, conversations with the agent, messagesUntil you delete them or your account. There's no automatic expiry.
Something you put in your library's trash30 days, then it's gone
Photos, videos, sounds and files you add to what you say to the agent30 days, unless the agent keeps one in your library (then as your library)
A message you delete, or a conversation you deleteGone from your side at once. The other person keeps theirs.
Sign-in codes10 minutes to use them. Used and expired ones are kept, as hashes with the address they went to; nothing removes them yet
Which days you use Flisk400 days
Server logs30 days
A stream's itemsAs long as that stream says: up to 30 days for your own
Guests (someone who only played, with no account)A year after they're last seen, with everything of theirs
ReportsThe copy of what was reported: until a person closes the report, then 12 months. The record of the report itself: no set limit
Database backups30 daily backups, and 7 days of its transaction log, on Google Cloud SQL
An inactive accountNothing removes it at present. The Terms let us close it after 24 months, and we'd email you first

Deleting

This is how to delete your Flisk account, or some of your data while keeping the account. Flisk is made by MakeMoreDigital Limited.

Deleting your account

It goes at once, with your screens that nobody else shares, tokens, your own key, your credit, rooms, library and its files, your own store, conversations with the agent, push tokens, follows, and the pictures of your rooms. Your conversations with other people go too, for them as well as for you: a conversation needs both people.

What outlasts it:

Deleting some of your data

You don't have to delete your account to delete some of your data:

Your rights

You can ask us for a copy of your data, to correct it, to delete it, to restrict or object to how we use it, and to withdraw consent. Write to hello@flisk.tv. You can also complain to the Information Commissioner's Office (ico.org.uk), or your own country's authority.

Children

Flisk is for people aged 13 and over. Where a phone can say how old someone is (above, Your age), the app asks it: under 13, it doesn't let them in, and 13 to 17 is treated as under 18. Otherwise we ask how old people are and take their word for it. There are no parental controls in Flisk itself.

Anyone who says they're under 18, or hasn't said yet, is treated more carefully, whatever their other settings say: they can only exchange messages and shared things with people they follow who also follow them, so there are no message requests from strangers; and adults can't find them by handle or contacts, or see their public playlists, unless they've followed that adult. This is a real protection only as far as people answer honestly. If you think a child under 13 is using Flisk, tell us and we'll remove the account.

Changes

We'll change this page when what we do changes, and tell you in advance by email or in the app if it matters.