Privacy Policy

Effective 2026-10-22

Kichma is an idea→ship lifecycle engine. This policy describes what we collect, why, who else processes it, and what we deliberately do not collect. It is written to be checked rather than skimmed — where a claim is about how the software behaves, the behaviour is the commitment.

If anything here is unclear or looks wrong, write to hello@kichma.ai.

Your account

To give you an account we store your name, email address and avatar as your identity provider reports them, plus the organisations and projects you belong to and your role in each.

Sign-in is by GitHub OAuth or an emailed link. There is no password, so we never store one. Sessions are records in our own database rather than signed tokens, which is what lets a removed user's access end immediately instead of when a token happens to expire.

What you write

The documents you write — ideas, requirements, specifications, plans and the tasks beneath them — are yours. We store them because storing them is the product, and we keep their history because an unchangeable record of what changed and who changed it is the thing Kichma exists to provide.

That history is append-only on purpose. Editing a document does not erase what it said before, and the audit trail records who acted, when, and what the value was on either side. This is a deliberate guarantee rather than a side-effect, and it means "delete this version" is not something the product can do for you. See Deletion for what we can do.

The agent, and what it sends

Kichma's interview and drafting agent is powered by Anthropic's Claude. When you ask it to draft, critique or restructure a document, the text of that document and your prompt are sent to Anthropic to produce the response.

🔴 This is the most consequential flow in the product and we would rather you knew it than discovered it: your product strategy leaves our infrastructure when you use the agent. It happens only when you invoke the agent — there is no background analysis of your content, and nothing is sent to a model to generate recommendations you did not ask for.

We do not use your content to train models, and Anthropic does not train on API content. If that is unacceptable for a particular project, the product is usable without the agent: every document can be written by hand.

Analytics, and first-touch capture

We measure how the public pages are used, and we do it on an analytics instance we host ourselves — no analytics vendor receives any data about you. There is no Google Analytics, no third-party tag and no advertising pixel on any Kichma page.

What is recorded for a visit: the page URL, the referring site, the browser and operating system family, approximate country, and screen size. The visitor's IP address is used to compute an identifier and is not stored. When you are signed in, your account id is attached so that usage can be understood per account rather than per anonymous session.

First-touch capture. When you create an account, we record how you first arrived — the referring site's hostname and any campaign parameters in the link you clicked — against your account, once. It is never overwritten, and it exists to answer which of our own efforts reached you. We store the referring hostname only, never the full URL, because the path belongs to someone else's site and may describe their users rather than ours.

Why there is no cookie banner

You will not see a cookie banner on this site, and that is not an oversight.

European law regulates storing or reading anything on your device, whether or not it identifies you. So we store nothing: our analytics sets no cookie, and first-touch capture works by carrying the campaign parameters along in the link you are already following, not by writing to your browser. No cookies, no localStorage, no device fingerprinting.

The only cookies Kichma sets are the ones that keep you signed in and remember your interface preferences. Those are strictly necessary to provide a service you asked for, which is why they need no consent — and why signing out removes them.

Billing

Payments are processed by Stripe. Card details are entered on Stripe's own checkout and never reach Kichma's servers; we store a Stripe customer reference, your plan and seat count, and your subscription status.

Once a day we copy a summary of paid invoices — the amount, the date, the plan and the acquiring campaign — into the analytics database so that usage and revenue can be looked at together. Those records name an organisation and an account id. They do not carry email addresses or payment details.

We process your data on these bases: to perform our contract with you (providing the service, billing), our legitimate interests (keeping the service secure and reliable, understanding how our own marketing performs), and legal obligation where one applies.

Where processing rests on legitimate interests — analytics and first-touch capture in particular — you can object by writing to us, and we will stop.

Your rights

You can ask us to show you what we hold about you, correct it, export it, or delete it. Write to hello@kichma.ai and we will respond within 30 days.

Export is also self-service: an organisation's admins can export its data in full from Settings → Export, on any plan and after cancelling, without asking us.

Deletion

There is no self-serve way to close an account yet. To have yours removed, write to hello@kichma.ai and we will tell you exactly what can and cannot be removed for it, and remove what can.

🔴 What cannot be removed is authorship. Because Kichma's promise is a record that cannot change without a trace, removing an account does not remove the documents it wrote or the history of changes it made. We are building account deletion that anonymises them instead — the entries will remain, attributed to a removed user rather than to you — and we will update this page when it ships. We are telling you this plainly because it is a real limit on erasure, it follows directly from what the product is for, and discovering it after the fact would be a breach of trust even where it is not a breach of law.

Organisation data is retained while the organisation exists. When a subscription ends, the workspace becomes read-only for at least 30 days — still readable and exportable — so that nobody loses a product record to a lapsed card. We do not currently delete content after that period, and we will update this page before that changes.

Security

All connections are encrypted in transit, and we verify the certificates of the databases we connect to. Access to production data is limited to those who need it. Every change to the record is attributed and retained.

If you believe you have found a vulnerability, write to hello@kichma.ai and we will respond.

Where data is processed

Kichma and its subprocessors are based in the United States and process data there. Where data is transferred from the European Economic Area or the United Kingdom, we rely on the Standard Contractual Clauses our subprocessors offer. Our Data Processing Addendum covers this for customers who need it in writing.

Changes to this policy

When this policy changes materially we will say so, and the date at the foot of this page is the one that counts. Older versions are in this product's own public commit history, which is a slightly unusual thing for a privacy policy to be able to claim and seems worth claiming.

Who else processes your data

Every third party that processes data on Kichma's behalf is named on the subprocessor list, with what each one receives and why. It is kept current by a test that holds it to the software's own dependency list, so a new service cannot be added without being disclosed.