What we store, and who else sees it.

The short version

Flummox stores your account details, the text you write, the reference images you upload and the media generated from them. To generate anything, that text and those images are sent to third-party model providers. We do not train models on your content and we do not sell it. Deleting your account deletes it.

One thing worth knowing up front: uploaded and generated files are served from public URLs. See Files are public by URL below.

Who we are

Flummox is operated by Implicit Marketing, Inc., a California corporation (“we”). We are the controller of the personal data described here. For anything in this policy — access, correction, export, deletion, or a complaint — write to support@flummox.ai. Postal address: 3017 Douglas Blvd #300, Roseville, CA 95661.

What is stored

Account. Your email address, your display name and avatar if you set one, which workspace you belong to and your role in it, and the timestamps of account events such as sign-up and sign-in. If you sign in with Google we receive your email address, name and profile picture from Google; we never receive your Google password.

Content. Project names and briefs; asset names, descriptions and the written likeness generated from an asset’s reference sheet; scene text including the tags linking a scene to an asset, and any camera or style direction attached to it; images you upload as references; images and video generated for you; and the render settings for each — model, duration, resolution, aspect ratio, whether audio was requested, and the seed.

Operational. Credit ledger entries (what a generation cost and when), render job records including the prompt actually submitted and any provider error message, and rate-limiting counters keyed to your workspace. Ordinary server logs, including IP address and browser user agent, are produced by our hosting provider for security and debugging; we do not export or keep a separate copy of them, and they are retained under that provider’s own policy.

Payment. Purchases are processed by Stripe. Your card details are entered on Stripe’s own forms and never touch our servers; what Flummox receives and keeps is your email, the amount, and a token identifying the transaction — the credit-ledger entry described above.

Files are public by URL

Reference images you upload and media generated for you are stored in public buckets and served from long, unguessable URLs. Anyone holding such a URL can open the file without signing in to Flummox.

This is not an oversight — the model providers fetch your reference images by URL in order to generate from them, and a link that expires would leave a render conditioned on a missing file. But it does mean the protection on your files is the secrecy of the link, not an access check. Treat a Flummox file URL the way you would treat an unlisted document link: sharing it shares the file.

Who it is sent to

These are the third parties the product depends on. All of them process data in the United States.

• Supabase — database, authentication and file storage, in the us-east-1 region. Holds your account record, all your content, and every uploaded and generated file.
• Vercel — hosting and serverless execution. Handles every request you make and produces the server logs described above.
• fal.ai — image and video generation. Receives the prompt for the scene or asset being generated, the reference images of the assets you tagged, any still you chose to animate, and the render settings. It does not receive your email, your name, or content from projects you did not reference.
• Anthropic — used to expand an asset description into per-view directions, to break a written idea into a shot list, and to write an asset’s likeness. It receives that text and, for the likeness feature, the reference images of that asset. It does not receive your email, your name or your billing details.
• Stripe — payment processing. Receives your email, billing details you enter on its checkout page (card number, billing address), and the purchase amount. Flummox never sees the card itself. Stripe’s own privacy policy governs what it keeps.
• Cloudflare — bot protection on the sign-in and sign-up forms. Receives your IP address and browser characteristics when those forms load, in order to decide whether you are a person. It receives no account content.
• Sentry — error monitoring. Receives the stack trace and message of server and browser errors. It is configured not to attach IP addresses, cookies or request bodies, session replay is disabled, and messages are scrubbed of anything shaped like a credential before they are sent.
• Google — only if you choose to sign in with Google, and only to authenticate you.

Each has its own terms and retention behaviour which we do not control. If you need the data processing agreements behind these, write to support@flummox.ai.

Training

We do not train any model on your prompts, uploads or output, and we do not sell or share your content for anyone else’s advertising.

On the providers: Anthropic publishes that it does not use data submitted through its API to train its models. For fal.ai we have not confirmed equivalent terms in writing, so we make no claim on their behalf — if that matters to your work, read their terms before uploading images of real people. We would rather leave this uncomfortable than reassure you about something we have not verified.

Cookies

Flummox sets the cookies needed to keep you signed in, and nothing else. There are no advertising or cross-site tracking scripts anywhere in the product.

We do count page views, using Vercel Web Analytics. It is first-party and cookieless — it stores nothing on your device and builds no profile of you across sites, which is why it needs no consent banner. It tells us which pages people open, not who opened them. If we ever add anything that does track you, this section changes with it and you will be asked first.

Images of people

A reference image of a person is sensitive: the whole point of this product is that it can be used to generate new pictures and video of them, and the written likeness we derive from it describes their appearance in detail. Only upload an image of someone if they have agreed to that, and read the Terms of Service on what you may make with it.

If you believe an image of you has been uploaded without your permission, write to support@flummox.ai. We will remove it, the reference sheet built from it, the written likeness derived from it, and the media generated from it.

Deleting your data

Deleting a project removes its assets, reference sheets, scenes and generated media. Deleting your account removes all of that, along with your profile and workspace. Account deletion and a full export are handled by request today — support@flummox.ai — and self-serve controls in Settings are planned.

Deleted rows may persist for a period in the encrypted backups our database provider takes on its own schedule; we do not keep separate backups of our own. Credit ledger entries may be retained in a form tied to the transaction rather than to you, where accounting and tax rules require a record of what was charged.

Your rights

Depending on where you live you may have the right to access, correct, export or erase your personal data, to object to some processing, and — for California residents under the CCPA — to know what is collected and to opt out of its sale. We do not sell personal data. We honour these requests from everyone, not only from people in jurisdictions that mandate them.

Write to support@flummox.ai. If you are in the EU or UK and are unsatisfied with our response, you may complain to your local supervisory authority.

Children

Flummox is for people aged 18 or over. We do not knowingly collect personal data from anyone under 18; if we learn that we have, we delete the account and its contents.

Changes

Material changes are announced in the product before they take effect, and the date at the top of this page moves. See also the Terms of Service.