Privacy Policy

What Nevzenit holds, why, who receives it and what you can ask for. Version 2026-11-12, in force from 12 November 2026.

On this page

Who is responsible for your data

Nevtool Ltd is the controller of the personal data described on this page. We are registered in England and Wales under company number 16393560, our registered office is 71–75 Shelton Street, Covent Garden, London WC2H 9JQ, United Kingdom, and we are registered with the UK Information Commissioner’s Office. Write to support@nevzenit.com about anything on this page.

When your projects contain personal data about other people, such as your clients, you decide how it is used and we process it for you. That is covered by the Data Processing Terms in section 15 of the Terms of Service.

What we collect, and why

  • Your account. Your email address, and your name and picture if you sign in with Google or Microsoft; the date you accepted these documents; your practice’s name, country and currency. We need these to provide the service under our contract with you.
  • Your content. What you put into projects: briefs, notes, messages, photographs, documents, logos and outputs. We use it only to run the features you use, under our contract with you.
  • Billing. Stripe, the Merchant of Record that sells your plan, tells us your plan, payment status and country. Stripe processes your payment details; we never see your full card number. We keep what tax law requires, which is a legal obligation.
  • Usage and security. Credit use, reason codes, product events and technical logs, used to run and secure the service, prevent abuse of free credits and fix faults. The lawful basis is our legitimate interest in a working, fair and secure service.
  • Messages to us. What you write to support, used to answer you.
  • Proposal acceptances. When your client accepts a proposal you shared, we keep, for you, the name they typed, the time, and an identifier of the exact version they were shown. For a link page we store no email address or IP address. When the client accepts inside a client room, we also keep the email address they signed in with, and for a fee variation the figure and change they were shown and the proposal it varies. We do not keep an IP address, user agent, device or location with it. It cannot be edited, and it is erased with the proposal, the project or your account. A client who wants their name or email corrected or removed asks you, because you decide what is done with it.
  • Shared links. When you share a proposal link, we record when it was first and last opened and how many times, so you can see that it arrived. We store nothing about the person or device that opened it.
  • Founding Architect invitations. If we invited you to join as a Founding Architect, we keep your first name, your office’s name and the note we wrote to you, so we can see what we sent, and the times the invitation page was first opened, you signed in from it, or you stopped it. We keep no email address with it. When you stop the invitations we erase your name, the note and those times, and keep only the invitation’s web address, which contains your office’s name, with the dates it was sent and stopped, so that we do not invite you again. If you have not signed up 12 months after the invitation, we erase its details in the same way. The lawful basis is our legitimate interest in inviting architects to the service.

We do not sell personal data, we do not share it for advertising, and we do not make decisions about you that have legal or similarly significant effects by automated means alone. We send service emails: sign-in links, notices about your account, your credits or a payment, and notices about these terms. These come whatever your settings say. We also send a few product emails about getting started with Nevzenit, on the basis of our legitimate interest and the UK rules for emails to existing customers. You can refuse them when you first set up your account, and stop them at any time with the link in every product email or in Settings. We send no other marketing email unless you ask for it.

Your work is private by default

Projects, briefs, project documents, messages and generated outputs are private to your workspace. Access is enforced twice, in server code and again by database row-level security, so a link or an identifier alone never grants access to anything.

Uploaded files and generated images live in private storage. There is no public bucket. The only thing anybody can open by link is a document you choose to send a client, such as a fee proposal.

Clients invited to a project

This section is for people an architect has invited to a project on Nevzenit. You do not have a Nevzenit account and you do not accept our Terms. The studio that invited you decides why and how your data is used (it is the controller). Nevzenit processes it on the studio’s behalf under the Data Processing Terms in section 15 of the Terms of Service, and the studio is who to ask first.

  • Your email address, and when you were invited, joined and, if it happens, had access removed.
  • What you did in the project: that you signed in, viewed or opened an item, accepted, chose an option, or sent a request, with the time. These records hold no email or message text.
  • What you write to the studio in a request, when you sent it, and whether the studio has seen it.
  • When you accept: the name you typed, the time, the email you signed in with, an identifier of the exact document you were shown, and for a fee variation the figure and change shown.
  • When you answer a question the studio asks you to decide: the option you chose, the options you were shown, the time and the email you signed in with.

We do not keep your IP address, your device or your location with your records in the project. Our hosting and sign-in providers keep short technical logs for security, which can include an IP address; they are not linked to what you see or write, and are deleted on their normal schedule. We do not send you marketing, we do not profile you, we do not sell or share your data for advertising, and what you write is never used to train AI. The studio sends you the invitation from its own email; we use your email address only to send your sign-in codes.

The studio uses AI inside Nevzenit to work on its project: to answer its own questions about the project and to prepare documents such as a fee proposal or a project summary. When it does that work, the name the studio holds for you, what you wrote to the studio in a request and what you chose or accepted in the project can be sent to the AI provider named below (Google Cloud), as part of the project, so that the result fits the person it is for. Your email address and your phone number are left out of what Nevzenit sends. The provider processes it only to produce that result for the studio, under terms that do not allow it to train its models on it. AI makes no decision about you: what it writes goes to the studio, which reads it and decides what to do with it. Nothing AI wrote reaches you unless the studio sends it. If you would rather your words were not used in this way, tell the studio; it can erase the text of a request, and an erased request is not sent.

Access stops as soon as the studio removes it; a file link you already opened works for at most five more minutes. Records of what you saw, accepted or asked stay with the project as the studio’s record and are erased when the studio deletes the project. Your sign-in account is deleted when your last invitation is removed or its project is deleted. Your email address stays in the project’s record of the invitation and of anything you accepted until the studio deletes the project. The studio can erase the text of a request earlier; a blank entry then remains showing it was sent. Database backups are overwritten on their normal schedule.

You can ask the studio to see, correct or delete your data. If you cannot reach them, write to support@nevzenit.com and we will pass your request on or act on it for them. You can also complain to the Information Commissioner’s Office (ico.org.uk) or the data protection authority where you live.

Model training

Your content is not used to train or improve Nevzenit’s models or any provider’s models. If that ever becomes possible, it will be a separate, specific and revocable opt-in that is off until you turn it on, never a term buried in an update.

What leaves Nevzenit, and what does not

Answering a question or producing an image means sending something to an AI provider. What is sent is what the request needs: the confirmed project facts, the relevant constraints, what you typed or pasted (up to 20,000 characters) and any photograph or drawing you attach, and never an entire project or an entire conversation. A PDF you attach to a Think question is read on your own device and only its text is sent: the pages a question needs, never the file itself, and never more of it than the request needs.

When a project has a client, what a request needs can include the client’s name as you recorded it, what the client wrote to you in the client room and what they chose or accepted there, so that an answer, a fee proposal or a summary fits the person it is for. You are not asked each time. Your client’s email address and phone number are left out of what Nevzenit sends, unless you type them into a question yourself. You are the controller of your client’s data (section 15 of the Terms of Service), so telling your client how you use it is your part; the section above, Clients invited to a project, is the notice a client is shown when they sign in, and you can point them to it.

The AI provider for text and images is Google Cloud (Vertex AI or the paid Gemini API), under terms that do not allow Google to train its models on your content. Payments are handled by Stripe, our Merchant of Record, which receives the transaction and never your project material.

Dictating into Think uses your browser’s own speech service (Google in Chrome, Apple in Safari), so the audio goes to that provider under its terms. Nevzenit receives only the text you then send, and never receives or stores the audio.

Logs and support

Logs and analytics carry opaque identifiers, reason codes and aggregate cost and latency figures. Prompts, images, documents, addresses and private project names are excluded from them by default, and signed URLs are short-lived and never written to a log.

Public pages are counted per page per day, with no address, no cookie and no account attached to the count. The count also keeps the country our host names for the request and one word for the kind of page a visit came from (a search engine, a social network, another site, this site, or none); the address of that page is not kept. Automated visits such as search crawlers and link previews are not counted. That count, and the handful of product events that would otherwise leave no record, are kept because they show where a visit stops. The lawful basis is legitimate interest. The public count is not personal data. A product event made while signed in is kept with the account it belongs to, at most once a day for each kind, and is erased with that account.

A support view shows metadata first. Opening private content takes a deliberate action that is recorded, and a rejected upload never lands in an ordinary support ticket.

Erasing your account

You can erase your account yourself, from Settings → Delete account. It is one action, it is not reversible, and it deletes your projects, project memory, briefs, thinking, visuals, uploaded images and every record of what you did. Workspaces you were the only member of go with them, and so does any link you sent a client: a proposal belongs to the project it was written for, so somebody holding its address afterwards is told the link is not available and nothing more.

What stays is the money: credits granted, credits spent and the settlements between them, with your name cut off the rows so they no longer point at a person. Work in a workspace you shared with somebody else stays theirs, and no longer carries your name.

The image files themselves are a separate request to the storage backend, made immediately after the erasure commits. The records that reach them are already gone by then and nothing can find the files through Nevzenit again, but if that request is interrupted or the backend refuses it, unreachable bytes can remain there until they are removed. This is written down rather than left out because it is the one part of erasure you cannot check from outside.

Taking a copy with you

You can download your own data yourself, from Settings → Download my data. It is one press and it produces one file: your account, the practices you belong to, your projects and their memory, your thinking, your visual work, the documents you have written and the record of what you spent. Nothing is stored to make it: the file is built while you wait and sent straight to you, so there is no copy of your account sitting anywhere afterwards and no link for anybody else to find.

What the file does not carry is the pictures themselves: uploaded photographs and generated images. Each one is named in it, and support@nevzenit.com sends the files when you ask.

Who receives your data

These sub-processors run parts of Nevzenit for us, each under a contract that binds them to protect the data and use it only for us:

  • Supabase: database, sign-in and file storage, in Frankfurt, Germany.
  • Vercel: hosting and delivery of the website and app. A United States company with servers in several countries.
  • Google Cloud (Vertex AI or the paid Gemini API): the AI models that answer questions, read documents and make images, under terms that do not allow Google to train on your content. The parts of a project a request needs are processed where Google serves the model, which can be outside the UK and Europe.
  • Error monitoring: reports of faults on our servers and in your browser, stripped of your content and your identity before they are sent. A report from your browser carries only the kind of fault, the app version, a random reference and a technical code: never what you typed, the page address or your account. The same service also receives two kinds of short notice that carry an account reference but no content: a request to join a workspace, and a payment dispute (its amount, currency and reason).
  • Cloudflare Turnstile: a check against automated abuse on the sign-in page, before an account is deleted, and before the free fee split calculator makes its PDF. Cloudflare processes your browser and connection details for that check, and nothing from your projects.
  • Resend: delivers the product and account emails described above. It receives your email address and the content of each email, and nothing from your projects. A United States company.
  • Google and Microsoft: only if you choose to sign in with them; they tell us your email address, name and picture.
  • Stripe, our Merchant of Record: the company that sells your plan, takes payment and handles tax. It receives what a purchase needs and never your project material, and it handles your payment details as its own controller, under its privacy policy at stripe.com/privacy.

We tell customers at least 30 days before adding or replacing a sub-processor that handles their content. We may also share data when the law requires it, to protect our rights or someone’s safety, or with a company that takes over Nevzenit, which would be bound by this policy.

Data outside the UK and Europe

Where personal data goes to a country without an adequacy decision under UK or EU law, the transfer is covered by the UK International Data Transfer Addendum or the EU Standard Contractual Clauses, or by the UK Extension to the EU–US Data Privacy Framework where the recipient is certified. You can ask us for a copy of the safeguards.

How long we keep it

Your account and content are kept while your account exists, and erased as described above when you erase it. Billing and tax records are kept for six years after the end of the tax year they belong to, because UK tax law requires it. Technical logs are kept for a short period, usually no longer than 30 days. Support messages are kept for two years after the conversation ends. The details of Founding Architect invitations are erased 12 months after they were sent if nobody signed up, or earlier when you stop them; the invitation’s web address and its dates stay, so that we do not invite the same office twice. Database backups are overwritten on their normal cycle.

Your rights

Under the UK GDPR, and the EU GDPR if you are in the European Economic Area, you can ask for a copy of your data, have it corrected or erased, restrict or object to how we use it, and take it with you. Most of this you can do yourself in Settings. For anything else, write to support@nevzenit.com; we answer within one month. You can complain to the Information Commissioner’s Office at ico.org.uk, or to the data protection authority where you live, though we would like the chance to put things right first.

The same requests are open to you wherever you are, including under the laws of the United States, Canada, Australia, New Zealand and Singapore, and you can complain to the privacy regulator in your country. We do not sell or share personal data for targeted advertising.

Cookies and storage in your browser

Nevzenit uses only what the service needs to work: a cookie that keeps you signed in, a short-lived cookie that shows a notice after an account is erased, and browser storage for things like an unsent draft or a collapsed panel. Browser storage stays on your device: erasing your account cannot reach it, so on a shared computer clear the site’s data in your browser. There are no advertising or analytics cookies, so there is nothing to consent to. Public pages are counted without a cookie, as described above.

Once a German version of the site is open, a small cookie called nz_site_lang remembers the language you chose (en or de) for one year, so the site does not redirect you every time. It holds no identifier and is never shared.

Children

Nevzenit is for professionals aged 18 or over, and we do not knowingly collect data about children as users. Take particular care with any image of a child in project material; the Terms of Service require consent for it.

Changes to this policy

When we change this policy in a way that matters, we tell you by email or in the product before the change applies and ask you to confirm you have read it. Each version is dated at the top of this page.

What the service does not cover

The data download carries your account, practices, projects, memory, thinking and documents, but not the image files, visual drafts, feedback reports or billing events. Until it does, write to support@nevzenit.com and we send them.

Back to top