Privacy Policy

The promises, plainly

Before the full policy, here is what it all comes down to. Every promise on this list is written out in exact terms further down, and every one of them is true.

• Here is everything we know about you: your name and your time zone. That's the whole of the personal profile the studio keeps, your name, because you're signed in with it, and your time zone, so the studio knows your local time. We don't build a picture of you beyond that. No behavioral profile, no tracking of what you do, nothing sold to anyone. (Your card details live with our payment processor, not with us, we never see them. Your work is stored so you can come back to it, but it isn't mined into a profile, see below.)
• Your work is yours. You bring a brief, the room does the work, and what it ships belongs to you. We don't claim it.
• You can take it the moment it exists. Every file the room makes is yours to download the instant it's there, including work from a brief you never finished, and work from a brief the room couldn't complete. Owning and downloading your work is never held back over money. It's yours outright.
• You own what you make. The intellectual property in your briefs and your delivered work is yours to keep, use, and sell. We don't take a stake in it.
• Every brief is its own sealed world. Each brief is worked entirely on its own, its own memory, its own life. Nothing carries from one brief to another: not between different people, and not even between your own briefs. A new brief starts fresh and knows nothing of the others. We never reuse one person's brief to serve someone else, and we never pattern-match ("we've seen a brief like this, we'll reuse the answer"). The one brief you return to is still that same brief, right where you left it, but separate briefs never share.
• We don't use your work, your briefs, or your information to train or improve any AI model. Not your work, not your briefs, not anything about you, and not to train or improve any model, ours or anyone else's. We store your work for one reason only: so you can come back to it. It waits in your room for your return, and that's all it's for. The provider that powers the room is contractually barred from training on your work too.
• It goes nowhere, and it stays. We don't read it in the normal course, we don't sell it, and we don't share it. It sits in your room so it's there when you come back, you can pick a brief up and resume it anytime. Nothing is deleted because a balance ran out; running low on funds pauses new work, it never touches what you already have.
• The one time we'd have to hand something over: a court order, and even then, narrowly, and we tell you. The law can compel us to disclose your data, a valid subpoena, court order, or warrant, or a genuine emergency where someone's at risk of serious harm. That's the only exception to "we don't read your work," and we treat it as narrowly as it deserves: we check that the request is valid, we push back on anything overbroad, we hand over only what the law actually compels, and wherever we're allowed to, we tell you before we do it, so you can protect your data. We never volunteer your work to anyone.
• Your memory is yours to keep or clear. What the studio learns about how you like to work stays with your account, and you can turn it off or wipe it whenever you want.
• We tell you what we keep and for how long. Delete a brief and all of it, delivered work included, leaves your room right away and our systems within 30 days. Delete your account and everything goes the same way. Backups take a little longer to cycle out, and we tell you the real window below, we don't round it down to sound better.
• Kaldune is used all over the world, and the promises follow you. Studio58, Inc. runs Kaldune from the United States, and people use it from everywhere. Wherever you are, these promises apply, and where your local law gives you rights over your data, we honor them.

That's the whole of it. The rest of this page is the same promises, in the exact terms a lawyer and a regulator would hold us to.

---

Where you are

Kaldune is used all over the world. Studio58, Inc. operates Kaldune from the United States, and people use it from many countries. Wherever you are, the promises above apply to you.

Some places give you specific legal rights over your data, and this policy spells the biggest ones out in full further down:

• If you're in the United Kingdom, the EEA, or Switzerland, the GDPR sections (§8 and §9) set out the legal bases we rely on and your rights in full.
• If you're in California, the CCPA/CPRA section (§9) does the same.
• If you're somewhere else, for example under Brazil's LGPD, Canada's PIPEDA, or another local privacy law, you may have similar rights, such as to see, correct, delete, or export your data. Where the law gives you those rights, we honor them.

Whoever and wherever you are, you can reach us at the contact below and we'll help. And wherever we send data across borders to run the service for you, we protect it with the transfer safeguards in §5.

---

The full policy

Who we are. Studio58, Inc. ("Kaldune," "we," "us," or "our") is the company that operates the Kaldune service. Studio58, Inc. is the data controller for the personal data described in this policy, it decides what we collect and why. When we say "the studio" or "Kaldune," we mean the Kaldune service; when we say "we," "us," or "our," we mean Studio58, Inc.

This policy explains what personal data we collect when you use Kaldune, why we collect it, who we share it with, how long we keep it, and the rights you have over it. It covers both individual customers and business customers, and it applies to Kaldune's users worldwide (see Where you are, above). Where a section applies only to one kind of customer or one place, we say so.

1. What we collect and why

We collect only what we need to run Kaldune for you, and we keep the personal profile we hold about you deliberately small. The plainest way to say it: the only personal data the studio holds about you as a person is your name and your time zone. Everything else on this page is either something we don't hold ourselves (your card details, held by our payment processor) or your own work (stored for you, not turned into a profile of you). Here is the full account.

Your personal profile, your name and your time zone. The studio keeps your name (because you sign in with it and we address you by it) and your time zone (read from your device's clock so the studio knows your local time). That is the entire profile we hold about you. We do not build any further picture of you, no behavioral profile, no record of your habits, no tracking of what you do across the service, and nothing about you is sold or shared for anyone else's purposes.

What you need to sign in. To run an account we also hold the credentials and settings the account itself needs, your email address (to sign in and to send you the service messages your account depends on), your password (stored hashed, never in the clear), and your language preference. These are the keys to your account, not a profile of you; we use them to let you in and to reach you about your account, and for nothing else. For business customers, this also includes the organization and the people it authorizes to use it.

Your card details, held by our payment processor, not by us. When you add funds, payment is handled by our payment processor, Stripe. The studio never sees or stores your full card details, they go directly to Stripe, which is bound to use them only to process your payment. See §4 (sub-processors) and, for the money terms themselves, the Terms.

Your work, stored for you, not studied. The briefs you write and the work the room delivers to you. We store this so your room is waiting for you when you return, so you can download it whenever you want, and so you can resume a brief where you left off. This is your content, and it is not part of the personal profile above, we don't read it in the normal course, we don't profile you from it, we don't sell or share it, and we don't use it to train or improve any AI model. See §2 (how the room works), §3 (every brief is sealed), §6 (retention), §7 (the one support exception), §5.3 (the one legal-disclosure exception), and §9 (your rights, including export).

Usage and technical data, kept to a minimum, and not about you. To keep the service running securely and to fix faults, we keep basic technical logs. We hold these to the aggregate level wherever we can, our error-diagnostics tooling is configured to scrub your work out of the records it collects, and we do not use them to build a profile of you or to track you across the service.

We do not collect special-category data (health, biometric, and the like) as a matter of course. If your briefs or your work happen to contain such data because that's the work you asked for, it is handled as your content under the promises above, not added to any profile of you.

2. How the room works, the AI provider

Kaldune's room is powered by Claude, an AI system made by Anthropic, which we access through Anthropic's commercial API. To answer your brief, the relevant parts of it are sent to Anthropic to generate a response, then returned to your room.

This matters for your privacy, so we are specific about it:

• We do not use your work, your briefs, or your information to train or improve any AI model of our own. Not to train a model, not to improve one, not for any model, ours or anyone else's. Your work is stored so you can come back to it, and for nothing else.
• Anthropic does not train its models on the work you send through the commercial API. This is a contractual commitment in Anthropic's Commercial Terms of Service, which state that "Anthropic may not train models on Customer Content from Services." We use the commercial API precisely so that this holds, Anthropic's separate consumer products behave differently, and we do not use them.
• Anthropic deletes what it processes. Anthropic automatically deletes the inputs and outputs it handles for us from its systems within 30 days.
• We surface no feedback control that would forward your work to Anthropic for training. Anthropic's one training exception is content a user voluntarily submits as feedback. Kaldune does not send your work to Anthropic as feedback, so this exception does not reach your work.
• The one exception neither we nor you control: if Anthropic's automated trust-and-safety systems flag content as violating Anthropic's usage rules, Anthropic may retain that content for up to 2 years and its safety-classification records for up to 7 years. We disclose this because it is true; it is not something we can switch off.

Anthropic acts as our data processor for this work. Anthropic in turn relies on its own infrastructure providers, Google Cloud Platform, Amazon Web Services, Microsoft Azure, and Cloudflare, to host and route data, worldwide. See §5 (international transfers) for how that is protected.

3. Every brief is sealed and worked on its own

Every brief in Kaldune is completely independent. Each one is worked in its own sealed context, its own working memory, and nothing carries from one brief to another:

• Nothing crosses between different people. We never use one person's brief, or the work in it, to serve another person. Your briefs are worked for you alone.
• Nothing crosses between your own briefs, either. A new brief starts fresh and knows nothing of your other briefs. The studio does not carry the content or context of one of your briefs into another one.
• We don't pattern-match across briefs. The studio does not look across briefs to say "we've seen a brief like this before, we'll give the same output." Every brief is worked on its own merits, from its own instructions.

How this fits with resuming a brief. A brief persists so you can leave it and come back, and the same brief, resumed, is still that one sealed world, right where you left it (see §6 and the Terms on ownership and resuming). Persisting and resuming happen within a single brief; they never merge one brief into another. Sealed-and-independent applies between briefs; resumable applies within one.

How this fits with search and memory. Two features touch more than one brief, and neither breaks the seal:

• Search lets you find your own briefs across your studio. That's you pointing a tool at your own studio to retrieve your own files, not the studio reading across your briefs, and not one brief's working context leaking into another. It is described the same way in Organizing your work and Trust & safety.
• Memory, the separate feature that learns how you like to work, is account-level and entirely under your control, you can turn it off or clear it anytime (see §9 and Privacy Choices). It holds preferences about how you work, never a store of your brief content carried from one brief into the next; and with it off, a new brief starts with nothing at all from any other.

4. Who we share your data with, sub-processors

We do not sell your personal data, and we do not share it except with the service providers we need to run Kaldune. Each is bound by contract to use your data only to provide its service to us, and nothing else. Our sub-processors are:

The providers marked (to be named) are confirmed and their locations published here before this policy goes live; each one's contract restricts your data to running Kaldune and nothing else. Stripe (payments) is confirmed and named, consistent with the Terms (§7.5) and the DPA. Anthropic's no-training commitment (§2) is stated absolutely because it is in Anthropic's Commercial Terms.

When we add a new sub-processor. We give you advance notice, our current policy is 30 days, before a new sub-processor starts handling your data, and business customers can object under the DPA.

5. International transfers

Kaldune operates from the United States, and some of our providers operate in the United States and other countries. When we move personal data out of the UK, the EEA, or Switzerland, we protect it with a recognized transfer mechanism:

• For transfers to our providers, we use the European Commission's Standard Contractual Clauses, Module Two (controller to processor) and Module Three (processor to processor down the chain), together with the UK International Data Transfer Addendum and the Swiss Addendum where they apply.
• Anthropic, as our processor, relies on the same Standard Contractual Clauses for the onward transfers to its own worldwide infrastructure providers (Google Cloud Platform, Amazon Web Services, Microsoft Azure, and Cloudflare).

If you are elsewhere in the world, your data is likewise processed in the United States and the other countries where our providers operate; where your local law requires a transfer safeguard, we apply one.

5.1 We don't sell or share your data

We do not sell your personal data, and we do not share your Customer Content, except with the sub-processors above who are contractually bound to use it only to run Kaldune for you. This holds worldwide.

5.2 We don't read your work in the normal course

In the ordinary operation of the service, no one at the studio reads your briefs or the work in your room. The only in-house exception is support, and only when you ask, that is set out in full in §7. Written here so this section is complete; the canonical wording is §7.

5.3 When a government or a court asks for your data

This is the one exception to §5.2 that neither we nor you choose, and we treat it as narrowly as it deserves. It is written here in the same words as the Terms (§5.3), so the two can never say different things.

When a government or a court asks for your data. We disclose your work or your personal data to a government or law-enforcement authority only when we're legally required to, when we're compelled by valid legal process, such as a subpoena, court order, or warrant, or in a genuine emergency where someone is at risk of serious physical harm. When we get a request like that, we don't just hand everything over: we check that it's valid, we push back on requests that are overbroad or vague, and we disclose only the specific data the law actually compels, nothing more. And wherever we're legally allowed to, we tell you before we disclose, so you have the chance to protect your data, unless a court has ordered us not to, or someone is in danger. A compelled disclosure like this is the one narrow exception to our promise that we don't read your work in the normal course (§7): the law can open this one door, and when it does, we open it as little as the law allows and we tell you we're doing it.

How this fits with the AI provider. Because Kaldune uses Anthropic's commercial API (not a consumer product), if a government seeks your content held at Anthropic's layer, Anthropic's published policy is to ask the requester to contact its customer, that's us, Studio58, Inc., in the first instance. So the request comes to us, where the narrow, notify-you-first posture above applies. Anthropic's separate consumer good-faith-sharing posture does not apply to Kaldune, because we are not on the consumer product. We never volunteer your work to law enforcement; we respond only to valid legal process.

6. Retention, how long we keep your work, and when it's gone

We keep two clocks, and we tell you both: our own storage (what Kaldune keeps so you can return to your work) and the AI provider's processing (what Anthropic holds briefly to answer your brief).

Two things this section does not do: it does not delete your work because a payment balance ran low or ran out, and it does not delete work from a brief you left unfinished. Your work leaves only when you delete it or close your account, or where the law requires it. Running low on funds pauses new work; it never removes what you already have.

Our storage:

A note on the backup window, kept honest. The "no longer than 90 days" above is our protective ceiling while we confirm the true rotation period of our backups. When it's confirmed, we publish the real number here, shorter if it's faster, and the true longer number if it's longer. We won't round it down to sound better. This is the one number on this page still being confirmed; it matches the same figure in the Privacy Choices page and the DPA, and all three change together.

The AI provider's processing:

Legal holds. We may keep data longer where the law requires it, to resolve a dispute, or to deal with misuse, at both layers above.

7. We don't read your work, and the one exception, stated plainly

This is the reconciliation of two things that are both true, written here exactly as it is written on our support pages so the two can never say different things:

We don't read your work, and here's the one exception, stated plainly.

In the normal course, no one at the studio reads your briefs or the work in your room. We're not sitting behind the glass reading over your shoulder.

The one exception is support. If you ask us for help with something in your work, and only then, a support specialist can open the specific brief or item you've pointed us to, for the specific purpose you asked about, and nothing more. Every one of those accesses is logged: who opened it, when, and why. You can ask us for that record.

We never browse your work for any other reason, and we never use it to train AI models.

There is one other exception, and it is not one we choose: when the law compels us to disclose your data. That is set out in §5.3, and it is the only other time your work leaves the normal course, narrow, checked, and told to you wherever we're allowed to tell you.

8. Legal bases for using your data (GDPR)

If you are in the UK, the EEA, or Switzerland, we rely on these legal bases under Article 6 of the GDPR:

Where we rely on legitimate interests, you can object, see §9.

9. Your rights

If you are in the UK, the EEA, or Switzerland (GDPR). You have the right to:

• access the personal data we hold about you;
• rectify data that's wrong or incomplete;
• erase your data ("the right to be forgotten"), subject to the legal-hold exceptions in §6;
• restrict or object to our processing, including processing based on legitimate interests and any direct marketing;
• data portability, receive your data in a portable format and have it sent onward where technically feasible;
• withdraw consent at any time where we rely on consent (for example, marketing), without affecting processing already carried out; and
• lodge a complaint with your data protection authority.

You can exercise the everyday versions of these rights yourself, memory and deleting a brief are controls in the product, and exporting your work or deleting your account is one email to privacy@kaldune.ai (see Privacy Choices for every control and its real mechanism). Exporting your work is always available while your account is open, and is not tied to your payment status. For anything else, contact us (§13). We respond within one month, as the GDPR requires.

If you are in California (CCPA/CPRA). You have the right to:

• know what personal information we collect, use, and disclose, and the categories of sources and recipients;
• access and receive a copy of your personal information;
• correct inaccurate personal information;
• delete your personal information, subject to the legal-hold exceptions in §6;
• opt out of the sale or sharing of your personal information, we do not sell or share your personal information as those terms are defined under the CPRA, so there is nothing to opt out of, but the right stands;
• limit the use of sensitive personal information, we do not use sensitive personal information for any purpose beyond providing the service; and
• not be discriminated against for exercising any of these rights.

You can exercise these rights through the controls and channels in Privacy Choices or by contacting us (§13). We do not use or disclose your personal information for cross-context behavioral advertising.

If you are elsewhere in the world. Kaldune serves users worldwide (see Where you are, near the top). Many countries, for example Brazil (under the LGPD, Lei nº 13.709/2018) and Canada (under PIPEDA), give you rights over your personal data that resemble the ones above, such as to confirm we hold your data, access it, correct it, delete it, or receive a copy. Where your local law gives you those rights and it applies to us, we honor them. Use the same contact in §13 and we'll help; we don't make you clear a higher bar because of where you live. We describe this honestly: we honor the data-rights your local law gives you where they apply, we don't claim a separately certified compliance program for every country's regime, and if you'd like to know how a specific right works for you, just ask.

Authorized agents and verification. We may need to verify your identity before acting on a request, and we honor requests made by an authorized agent on your behalf where the law provides for it.

10. Cookies and tracking

Kaldune uses only the cookies it needs to work, the session and sign-in cookies that keep you logged in. For understanding aggregate usage, our default is privacy-respecting analytics that don't set tracking cookies and don't identify you individually, so there's no consent banner to click through.

Marketing emails, if you opt in, always carry an unsubscribe link, and we don't sell your contact details.

11. Security

We protect your data with encryption in transit and at rest, access controls that give staff only the access they need, and the operational practices of a service that takes this seriously. Our AI provider, Anthropic, maintains its own security program, encryption (AES-256 at rest, TLS 1.2 or higher in transit), least-privilege access, annual third-party audits and penetration testing, and independent certifications (SOC 2, ISO 27001). No system is perfectly secure, but we hold ourselves and our providers to a high, verifiable standard.

If a data breach affects your personal data, we notify you and the relevant authorities as the law requires, without undue delay.

12. Changes to this policy

When we change this policy, we update the "last updated" line and the version at the top. If a change materially affects your rights, we'll tell you directly, by email or in the product, before it takes effect, and, where the law requires it, we'll ask for your agreement.

13. Contact us

For any privacy question, to exercise a right (wherever in the world you are), or to ask for the record of any support access to your work, contact us:

• By email: privacy@kaldune.ai
• The data controller: Studio58, Inc.

Business customers processing their own end-users' personal data through Kaldune should also see our Data Processing Addendum, which governs that relationship and in which Studio58, Inc. acts as your processor.