Privacy notice
What we collect when you use daidocs.com, why we collect it, how long we keep it, and what you can make us do about it. Written to be read, not to be survived.
Last updated 19 August 2026 · Applies to daidocs.com and the Kerneta API
What it comes to
The product is built so that the sensitive material never reaches us. Your .dai files live on your machine. Your model provider key is held in your browser tab and forwarded per request, and not written to our storage. If you would rather not paste it each time, you can choose to save one to reuse, and then we keep it encrypted. The website runs analytics and advertising tags only if you agree to them, and nothing beyond the cookie recording that choice runs before you do.
- If you only read the site, we hold your IP address and request details in a web server log, and nothing else.
- If you send us an enquiry, we hold what you typed into the form until the conversation is finished.
- If you create an account, we hold your email, your name and a hash of your password, plus a record of what you converted so we can count your free conversions and bill a plan.
- We do not hold your model provider key unless you choose to save one to reuse, and then only encrypted, and we never sell anything to anyone.
Who we are
daidocs.com and Kerneta are trading names of Siro Robotics Ltd, a private limited company registered in England and Wales.
- Company number 10631734
- Registered office 124 City Road, London, England, EC1V 2NX
- Contact for privacy matters privacy@kerneta.com
Siro Robotics Ltd is the controller for the personal data described here: we decide what is collected and why. We have not appointed a Data Protection Officer, because we are not required to. Privacy questions go to the address above and reach a person, not a queue.
What we collect
Grouped by what you did to cause it, because that is the only grouping that lets you predict what we hold about you.
Reading the site
Our web server records the request: your IP address, the page you asked for, the time, your browser's user agent string and the page that referred you. This is ordinary server logging and it happens before any code of ours runs. It is how we find faults and see attacks.
Sending an enquiry
The investor form on the invest page collects your name, email address, fund or company, investor type, indicative amount and timing, and your message. The general enquiry tab of the same form collects your name, email, company or project, topic and message. The supporter form on the community page collects the name or handle you choose to give, an email address, and optionally a link to your site or profile.
You choose every field. Nothing is inferred, enriched or bought in from a data broker.
Creating an account
We collect your email address, the name you give, and your password. The password is stored only as an scrypt hash with a per-account salt: we cannot read it, and neither can anyone who steals the database. If you sign in with Google or GitHub instead, we receive your email address and display name from them and never see a password at all.
In use, we record which conversions you ran and their size, so that we can count the three free conversions, bill a plan, and enforce rate limits. We record the API keys we issue to you, and when they were last used.
To keep the free tier from being farmed, we also record a device identifier alongside your IP address when you run a free conversion. It is not a name or a serial number: it is a one-way hash of a few ordinary characteristics your browser reports anyway, namely its type, platform, screen size, timezone, language and processor count, with no reading of hardware, no canvas or font probing, and no cross-site tracking. We use it only to tell whether the same device has already used its free conversions. It also lets anyone refused unfairly, since a shared office or a family computer looks the same to it, ask us to lift the block. If you pay, or bring your own model key, none of this applies.
Why, and on what basis
UK GDPR requires a lawful basis for each purpose. Ours:
- Running your account and the API: performance of a contract. You asked us to provide the service; we cannot do it without an identifier and a usage record.
- Answering your enquiry: legitimate interests. You wrote to us wanting a reply. Our interest in replying does not override your rights, because replying is the outcome you asked for.
- Server logs, abuse prevention and security: legitimate interests. Keeping the service up and unbreached is in every user's interest as well as ours.
- Billing records: legal obligation. Company and tax law requires us to keep financial records for six years.
- Anything optional, such as a mailing list: consent. Asked for separately, refusable without consequence, and withdrawable at any time.
We do not carry out profiling and we make no automated decisions that produce legal or similarly significant effects about you.
Your model keys
This one is worth stating plainly, because it is the question a careful reader asks first.
When you use the converter with your own Anthropic or OpenAI key, that key is held in your browser's sessionStorage, which the browser empties when the tab closes. It is attached to each request as a header, used to call the model provider on your behalf, and discarded. It is never written to our database, never logged, and never visible in our admin tooling. If you would rather it never traverse our servers at all, self host: the engine is Apache 2.0 and the install guide covers running it entirely on your own machine.
Content you convert
A chat history you convert is your material and frequently personal, sometimes about people who are not you. Treat it accordingly, and so will we.
- We process it to perform the conversion you asked for, and for no other purpose.
- We do not use it to train models. Not ours, not anyone's.
- It passes to the model provider you selected so the conversion can run. See who else sees it.
- You can delete a store at any time, and deletion removes the content from our systems within 30 days including backups.
If the history contains personal data about other people, you are the controller of that material and we process it as your processor, on your instructions.
Cookies and analytics
One cookie is present before you answer anything: cookieyes-consent, set by CookieYes, our consent manager. It records the answer you gave the banner. It is strictly necessary in the sense the law uses: without it we would have to ask you again on every page.
Everything else waits for that answer. Google Analytics, Microsoft Clarity and the Meta pixel are delivered through Google Tag Manager, and CookieYes holds each of them back until you have agreed to the category it belongs to. Refusing is one click, in the same styling as accepting, and closing the banner without answering is not agreement.
Your dashboard at dashboard.daidocs.com sets one cookie of its own, kerneta_session, which keeps you signed in. It is marked HttpOnly, so no script running on the page can read it, and SameSite=Lax, so your browser does not send it from other sites. It is strictly necessary to keep you signed in, which under PECR means no consent is required for it, and it is never used to track you.
You can change your mind at any time from Cookie preferences in the footer of any page. Full detail is on the cookie policy.
Who else sees it
We sell nothing and we share nothing for advertising. Data reaches these third parties only:
- The model provider you choose (Anthropic, OpenAI or Google). Content you convert is sent to them to perform the conversion, under their own terms, using your key or, on a Claude subscription, the session you are already in.
- Our hosting provider, which operates the server the site and API run on and therefore holds the logs and database at rest.
- Stripe, our payment processor, when you subscribe to a paid plan. You enter your card on Stripe’s own hosted page and Stripe handles it directly. We never see or store a card number, only the customer and subscription references Stripe gives us so we know which plan you are on.
- Professional advisers, or a public authority, where we are legally required to disclose. We will tell you unless we are prohibited from doing so.
- CookieYes, Google, Microsoft and Meta. CookieYes is our consent manager and loads on every visit, because asking you is the one thing that cannot wait for your answer. The rest load only if you agree: Google Analytics and Microsoft Clarity for analytics, the Meta pixel for advertising. All arrive through Google Tag Manager. They receive your IP address, the pages you view and how you interact with them. Refuse, and none of them loads at all. See the cookie policy for the full list.
Those three are the only recipients that exist because we chose them rather than because the service needs them, which is why they are the only ones you are asked about.
Transfers outside the UK
Anthropic and OpenAI process data in the United States. Where content you convert is sent to them, it leaves the UK. The same is true of Google, Microsoft and Meta if you accept analytics or advertising. Those transfers rely on the UK International Data Transfer Addendum to the EU Standard Contractual Clauses, together with each provider's own published safeguards.
If that matters to you, it is avoidable rather than unavoidable: self host the engine with your own key or your own subscription, and no content reaches us at all.
How long we keep it
- Server logs: 90 days, then deleted.
- Enquiries: 24 months from the last message, so we can pick up a conversation you resume.
- Account data: for as long as the account exists, then 30 days after you delete it.
- Converted content and stores: until you delete them, then 30 days including backups.
- Billing records: six years from the end of the accounting period, because tax law says so.
How it is protected
Passwords are stored as scrypt hashes with per-account salts, which is a deliberately slow function chosen to make an offline attack on a stolen database expensive. The whole site is served over HTTPS with certificates that renew automatically, and every other domain we own redirects to it rather than serving anything. Model keys are not persisted server side unless you choose to save one to reuse, in which case they are stored encrypted. Access to production is restricted to the founders.
No system is perfect, and we would rather say that than imply otherwise. If you believe you have found a vulnerability, please write to security@kerneta.com and we will work with you.
Your rights
Under UK GDPR you can ask us to do all of the following, free of charge, and we will respond within one month:
- Access. Get a copy of everything we hold about you.
- Rectification. Have anything wrong corrected.
- Erasure. Have it deleted, where we have no overriding obligation to keep it.
- Restriction. Have us stop processing it while a dispute is resolved.
- Portability. Receive it in a machine readable format, or have it sent onward.
- Objection. Object to processing we base on legitimate interests.
- Withdraw consent, where consent was the basis, without affecting what was lawful beforehand.
Write to privacy@kerneta.com. We may ask you to confirm your identity, purely so we do not hand your data to somebody else.
Complaining
If we have handled your data badly, tell us first and we will try to put it right. You also have an unconditional right to complain to the supervisory authority:
Information Commissioner's Office, Wycliffe House, Water Lane, Wilmslow, Cheshire, SK9 5AF. Helpline 0303 123 1113, or ico.org.uk/make-a-complaint.
Changes
When this notice changes we will update the date at the top. If a change materially affects what we do with data we already hold, we will tell account holders by email before it takes effect, rather than relying on you to re-read the page.
Contact
Privacy questions and rights requests: privacy@kerneta.com. Anything else: hello@kerneta.com. By post: Siro Robotics Ltd, 124 City Road, London, England, EC1V 2NX.
Related: Terms of service · Cookie policy · Software licence