In plain terms
- We collect what we need to run your account: your email address, your name, your password (stored only as a one-way hash) and how you sign in.
- What you send to the API, such as models, reference rows, logged predictions and explanations, is your data. We process it only to provide the service to you. We do not sell it, use it for advertising or train AI models on it.
- We never receive your original model file, and card details go to Stripe, never to us.
- There is no advertising or cross-site tracking on our websites, which set no cookies of their own and count page views only if you allow it. The dashboard and API set no cookies of their own either.
- Email [email protected] to see, correct, export or delete your personal data.
This summary is here to help you read the policy. The full text below is what applies.
01Who we are and what this policy covers
Kaimei AI, Inc. (“Kaimei”, “we”, “us”) provides an explainability and monitoring service for machine-learning models. This policy explains how we handle personal data when you:
- visit our websites, including kaimei.ai;
- create or use an account on the dashboard at dashboard.kaimei.ai;
- call the API at api.kaimei.ai, directly or through the
kaimeiPython client; or - contact us.
It does not cover a Kaimei API that you or your organisation run on your own infrastructure. A self-hosted deployment sends nothing to us, and whoever operates it is responsible for the data in it.
Our Terms of Use govern your use of the service. If your organisation has signed an agreement with us that covers data protection, that agreement applies where it differs from this policy.
02Two kinds of data
We handle two kinds of data, and our role is different for each.
- Account data is information about you as a user of Kaimei: who you are, how you sign in and how you pay. We decide how it is used, so for this data we are the controller (under the GDPR and UK GDPR) or the business (under US state privacy laws).
- Customer Data is what you or your organisation send to the service to be analysed: models, feature schemas, reference rows, logged predictions, and the explanations and analyses we produce from them. We process it on your behalf and on your instructions, as your processor or service provider.
If Customer Data contains personal data about people your model makes decisions about, your organisation is responsible for that data, and those people should contact your organisation about it. Section 4 explains how we handle Customer Data.
03Account data we collect
What you give us
- Your email address, name and password when you sign up. We store your password only as an Argon2 hash, so we cannot see or recover it.
- A profile photo, if you add one. You can remove it at any time.
- Organisation details: the name of a workspace or team you create, the email addresses of people you invite, and each member’s role.
- Messages you send us. The contact form on our website opens your own email application, so what you write reaches us as an ordinary email.
When you sign in with Google or GitHub
We receive your account identifier at that provider, your email address, your name and the address of your profile picture. We ask only for basic profile access (from Google: openid, email and profile; from GitHub: read:user and user:email). We do not receive your password at that provider or any access to your repositories, files or contacts.
When you buy a paid plan
Payments are handled by Stripe. We send Stripe the email address of the person paying, the workspace name and our internal identifiers for it. Stripe collects card details on pages it hosts and, where tax applies, your billing address and tax ID. Card numbers never reach Kaimei. We keep the Stripe customer and subscription identifiers, your plan, its status and when the current period ends.
Created as you use the service
- When you signed up, when you last signed in, and whether your email address is verified.
- For each API key: its name, a hash of the key, who created it and when it was last used. The key itself is shown once and cannot be recovered from what we store.
- Security records. To limit sign-up, sign-in and password-reset attempts we record your IP address, and delete it after about two hours. When you sign up, resend a verification email or reset a password, Cloudflare Turnstile checks that you are not an automated program, and Cloudflare receives your IP address to do so.
- Server logs. Our hosting providers keep short-lived logs of requests (IP address, time, the path requested and the response status) so we can run and secure the service. We do not log request bodies.
On our websites
- We show no advertising and do no cross-site tracking.
- kaimei.ai is delivered through Cloudflare, which may set cookies that are strictly necessary for security.
- If you allow analytics in Cookie Settings, Cloudflare Web Analytics gives us aggregate page-view and page-speed statistics without cookies. It stays off until you allow it.
- Our pages load fonts from Google Fonts, and some load scripts from public content delivery networks, so your browser sends your IP address and browser details to those providers when it fetches them.
- Your cookie choices are saved in your browser, and so, unless you switch off Preferences in Cookie Settings, are your light or dark theme and your settings in the docs. None of them is sent to us.
04Customer Data
What we receive depends on how you connect a model
| Data | When we receive it |
|---|---|
| Feature schema: input names, types and outputs | For every model you register |
| Model structure and learned weights | When you upload a model with the Python client’s default settings. With weights="summary" we receive only shapes and statistics. |
| Reference rows and their labels | Only if you send them, and at most 5,000 rows per model |
| Logged predictions: inputs, outputs, your own ids and the outcomes you report | When your systems log them |
| Explanations, analyses and insights | When you or the service create them |
| A scoring endpoint’s address and the headers it needs | If you register one. Headers are encrypted at rest and never returned by the API. |
We never receive your original model file (the pickle, .h5 or .pt), and no part of your training set beyond the reference rows you choose to send. Model uploads also carry descriptive metadata: the client, Python and framework versions, the model’s class and the first line of its documentation. If you register a scoring endpoint, the service sends rows to it to obtain predictions, as you instruct. The Security page shows what crosses your network in each mode.
How we use it
Only to provide the service you asked for: storing it, computing explanations, analyses and drift checks, showing it in the dashboard and returning it through the API. We do not sell Customer Data, use it for advertising or use it to train AI models, ours or anyone else’s. We do not look at it except when you ask us for help with it, when it is necessary to keep the service secure, or when the law requires it.
Personal data inside Customer Data
Model inputs can describe people: a loan applicant, a patient, a subscriber. Send only the inputs your model needs, without names, account numbers or other direct identifiers, and de-identify health data wherever you can. Do not send protected health information as defined by HIPAA unless we have signed a business associate agreement with you.
Keeping and deleting it
Customer Data is kept until you delete it. Deleting a model permanently deletes its predictions, explanations, analyses, insights and uploaded structure at once, and deleting a team organisation deletes everything in it. Deletions cannot be undone. Copies in our database provider’s backups expire on that provider’s backup schedule.
05How we use account data, and why
We use account data for the purposes below. Where the GDPR or UK GDPR applies, the second column gives the legal basis we rely on.
| Purpose | Legal basis |
|---|---|
| Create and run your account, sign you in and provide the service | Performance of our contract with you |
| Take payments and keep invoices and tax records | Contract; legal obligation |
| Send service email: verification, password reset, invitations, and billing and security notices | Contract; legitimate interests |
| Protect the service with rate limits and bot checks, prevent abuse and fix problems | Legitimate interests in keeping the service secure and working |
| Understand, in aggregate, how our websites are used | Your consent, which you can withdraw at any time in Cookie Settings |
| Answer your messages | Legitimate interests; contract where you are a customer |
| Comply with the law and enforce our terms | Legal obligation; legitimate interests |
We do not send marketing email today. If we start, we will ask for your consent where the law requires it, and every message will let you unsubscribe. We do not make decisions about you that have legal or similarly significant effects by automated means.
07International transfers
Kaimei AI, Inc. is a US company, and our service providers process data in the United States and other countries. When personal data protected by the GDPR or UK GDPR is transferred out of the European Economic Area or the United Kingdom, we rely on a lawful transfer mechanism: the European Commission’s standard contractual clauses and the UK addendum to them, or a provider’s certification under the EU-US Data Privacy Framework and its UK extension.
08How long we keep it
| Data | How long |
|---|---|
| Account data | While your account exists. When you ask us to close your account, we delete your personal data within 30 days, except what the law requires us to keep. |
| Customer Data | Until you delete it (see section 4) |
| Billing and tax records | As long as tax and accounting law requires, usually up to seven years |
| Rate-limit records with IP addresses | About two hours |
| Server logs | A short period set by our hosting providers |
| Revoked API keys and expired invitations | As long as the workspace exists, without the secret, so that a revoked key stays revoked |
| Email correspondence | As long as we need it to deal with your message and anything that follows from it |
09How we protect it
- Every connection to the service is encrypted with HTTPS, and the API connects to its database over an encrypted connection.
- Passwords are stored as Argon2 hashes and API keys as SHA-256 hashes. Neither can be recovered from what we store.
- The headers a scoring endpoint needs are encrypted at rest and never returned by the API.
- The API will not call private network addresses on your behalf.
- We limit access to production systems to the people who need it to run the service.
No way of storing or sending data is completely secure. If a breach affects your personal data, we will tell you, and the authorities, as the law requires.
10Your rights and choices
In the product
You can change your name and photo in the dashboard, revoke API keys, leave or delete organisations, and delete models, projects and their data in the dashboard or through the API, which also lets you export them.
By request
Email [email protected] to access, correct, export or delete your personal data, including to close your account. We will check that the request comes from you, usually by replying to the email address on your account, and answer within one month or any shorter period your local law sets.
In the EEA and the UK
You have the right to access your personal data, to have it corrected or erased, to restrict or object to our processing of it (including processing based on legitimate interests), to receive it in a portable form, and to withdraw consent where we rely on it. You can also complain to your local data protection authority, such as the Information Commissioner’s Office in the UK.
In California and other US states
Depending on where you live, you may have the right to know what personal information we hold about you, to access, correct and delete it, to receive a portable copy, and to opt out of its sale, of targeted advertising and of profiling. We do not sell personal information or share it for targeted advertising, and we honour Global Privacy Control signals. You may use an authorised agent to make a request, and we will not treat you differently for exercising your rights.
In the past 12 months we have collected these categories of personal information: identifiers (such as your name, email address, IP address and account identifiers); commercial information (your plan and purchase history); internet or other network activity (sign-in times, server logs and website statistics); and professional information you choose to give us, such as your organisation’s name. We collect it from you, from Google or GitHub when you sign in with them, and from Stripe. We use it for the purposes in section 5 and disclose it, for business purposes only, to the service providers in section 6. The only sensitive personal information we collect is your sign-in credentials, which we use only to sign you in.
If an organisation sent us your data
If your data reached Kaimei because an organisation used it with one of its models, for example as an applicant scored by a lender’s model, please contact that organisation. We will help our customers respond to your request.
12Children
Kaimei is a professional service for people aged 18 and over, and is not directed at children. We do not knowingly collect personal data from children. If you believe a child has given us personal data, tell us at [email protected] and we will delete it.
13Changes to this policy
When we change this policy we will update the date at the top of this page. If a change materially affects how we handle your personal data, we will tell you by email or in the dashboard before it takes effect.
14Contact
Kaimei AI, Inc. is responsible for the personal data described in this policy. For questions about it, or to exercise your rights, email [email protected]. For questions about our Terms of Use, email [email protected].