Privacy Policy

pursuant to Articles 13 and 14 of Regulation (EU) 2016/679 (GDPR)Last updated: 30 October 2026

1. Data controller

The controller of your personal data is Alessandro Ceruti, an individual who runs Bloomwright as a personal project.

Contact email: liebyrinth.support@gmail.com

The controller's postal address is available on request by writing to the email address above.

2. Personal data we collect

Bloomwright collects and processes the following personal data:

Email address — provided by the authentication provider (Google) when you sign in.
Google account identifier (provider account ID) — used to link your Bloomwright account to the authentication provider.
Display name and profile picture — the name and profile image on your Google account, received from Google when you sign in and refreshed from Google every time you sign in again. They are used to identify your account in the administration views the operator uses to run the service, and your display name is used as the author name on the books you export.
Project content — the data you enter or generate through Bloomwright (project sheet, chapters, scenes, characters, settings, story structure, conversations with the AI agents). This content stays associated with your account.
Service usage data — available credits, credit history, and the cost of the API calls associated with your account.
Technical data — the date your account was created and last updated.
Country of signup — the country your connection was coming from the first time we recorded it; for a new account, that is when you create it. It is derived from your IP address by our hosting provider and stored as a two-letter country code. Your IP address itself is not stored on your account record or elsewhere in our database.
Signup source — if you reached Bloomwright through a link that carried a campaign or referral tag, that tag is recorded once, at signup, so we can tell which channels bring writers to the service. It contains no information about you.
Arrival page — the page you were on when you signed up, and the host of the site that sent you there (for example google.com — never the full address), so we can tell which of our pages bring writers to the service. It is resolved once, at signup, from the page views of the visit you signed up in; the random session identifier that links them is used for that one lookup and is not added to your account.
Invitation used — if you joined through an invitation link we sent, we record which invitation your account accepted. It tells us the invitation was taken up; it says nothing else about you, and the link is removed if your account is deleted.
Poll answers — the answers given voluntarily in the in-app polls, including stated price preferences, kept linked to the account because completing one grants a usage bonus that has to be accounted for.
Feedback — the free text sent through the feedback page, linked to the account so we can reply to it and so it is deleted with the account.

We do not collect your physical address or your phone number, and we collect no personal data beyond what is listed above.

3. Purposes and legal basis for processing

The data we collect is processed for the following purposes, each with its own legal basis:

Purpose	Legal basis	Data involved
Authentication and account management	Performance of a contract (Art. 6(1)(b) GDPR) — necessary to provide the service you requested	Email, provider account ID, technical data
Running the credit system	Performance of a contract (Art. 6(1)(b) GDPR)	Usage data
Operating the AI agents (assisted prose writing, structure, characters, critique)	Performance of a contract (Art. 6(1)(b) GDPR)	Project content sent to the AI providers
Storing and editing your project	Performance of a contract (Art. 6(1)(b) GDPR)	Project content
Linking an invitation we sent to the account that accepted it	Legitimate interest (Art. 6(1)(f) GDPR) — our interest in knowing whether the invitations we send are taken up, weighed against a minimal impact on you: the link exists only because you chose to use an invitation, it is visible only to the operator, and it is removed if your account is deleted. You may object at any time (Art. 21).	The invitation used and the account that accepted it
Measuring page views (how many people reach Bloomwright, and which pages they read)	Legitimate interest (Art. 6(1)(f) GDPR) — our interest in knowing whether the site works and is found, weighed against a minimal impact on you: no profiling, no cross-site tracking, no persistent identifier, no IP retention. You may object at any time (Art. 21).	Page path, host of the referring site, a random session identifier
Arrival page — recording the page you arrived on before signing up	Legitimate interest (Art. 6(1)(f) GDPR) — our interest in knowing which of our pages bring writers to the service, weighed against a minimal impact on you: it is one page path and one referring host, recorded once at signup and never updated, visible only to the operator, and the session identifier used to find them is not added to your account, so no browsing history is linked to your account. You may object at any time (Art. 21).	Page path and host of the referring site, recorded once on your account

4. Recipients and data sharing

Personal data is never sold or transferred to third parties for marketing or profiling purposes. Data may be disclosed to the following parties, solely for the purposes stated:

Google Ireland Ltd — for the OAuth authentication process. Google acts as an independent controller for its own processing. See Google’s privacy policy.
Google LLC — project content and agent conversations are processed by the Gemini artificial-intelligence models to generate responses and prose. The content sent comprises the text of your project (chapters, scenes, characters, structure) and includes no identifying personal data other than whatever you may have written into that text yourself. Bloomwright uses the paid Gemini API (Google AI/Cloud "paid services" tier): under Google’s terms of service, content sent through this tier is not used to train Google’s models, nor shared with other organisations.
Anthropic PBC — part of the planning and review work runs on Anthropic's Claude models; Anthropic is also the fallback provider when a call to another model fails, whether that call went to Gemini or to the DeepSeek models. The content sent is the project text described above. Under Anthropic's commercial terms, inputs and outputs sent through the API are not used to train its models.
OpenRouter, Inc. — prose generation, the editor's chat assistant, scene planning, the onboarding conversation and four background helpers — condensing a worldbuilding note, breaking a scene into steps, checking the shape of a turning point, and reading a manuscript you import — are routed through OpenRouter to the DeepSeek models. For prose the content sent is the project text described above; for the chat assistant it also includes your story bible, a structured summary of your project (chapters, scenes, characters, settings, structure) and the conversation you are having with it; for scene planning it is your story bible and the kind of story it is, the turning point being planned, the scenes already planned under it, your chapter titles, your characters and settings, and any steering you type; for the onboarding conversation it is what you type into the public onboarding wizard — the idea you start from, your answers to its questions and the conversation so far — which is sent in order to answer you, including when you use the wizard before you create an account; for the background helpers it is the note's title, tags and text with an excerpt of your story bible; your story bible with the scene's chapter, summary, goals, characters, setting and existing steps; a turning point's title and summary; and, when you import a manuscript, its full text, sent one section at a time. Bloomwright restricts this routing to endpoints OpenRouter designates zero-data-retention — meaning the inference provider does not retain the prompt or the generated text once the request completes — and pins it to a fixed allowlist of inference providers acting as sub-processors: Novita, IonStream and DeepInfra for prose, and DeepInfra alone for the chat assistant, for scene planning, for the onboarding conversation and for the background helpers. The non-zero-data-retention direct endpoint is excluded.
Neon, Inc. — for storing account data and project content on a managed PostgreSQL database.
Upstash, Inc. — for technical rate limiting of requests (serverless Redis). Only technical request identifiers (IP address) are transmitted, for the purpose of protecting the service.
Vercel, Inc. — for hosting and delivering the application.
Resend, Inc. — for delivering operational notification emails to the operator of the service, not to you: a new signup, a feedback submission, and alerts about the platform budget. The data transmitted is the email address and the display name on the account, and the text of any feedback you chose to send.

5. Transfers to third countries

Account data and project content are stored on servers located in the European Union or the United Kingdom (depending on the configured Neon region).

Some service providers are based in the United States (Google LLC, Anthropic PBC, OpenRouter Inc., Vercel Inc., Upstash Inc., Neon Inc., Resend Inc.). Transfers of data to the United States take place on the basis of the EU-U.S. Data Privacy Framework (European Commission adequacy decision of 10 July 2023, under Art. 45 GDPR) where the provider is certified, or otherwise under Standard Contractual Clauses (the model clauses approved by the European Commission, under Art. 46(2)(c) GDPR).

Prose generation may be served by any of the inference providers OpenRouter is allowed to route to (Novita, IonStream, DeepInfra); the editor’s chat assistant, scene planning, the onboarding conversation and the background helpers are routed to DeepInfra only. Some of these are established outside the European Economic Area. Those transfers rely on the same safeguards — an adequacy decision where the provider is covered by one, otherwise Standard Contractual Clauses — and the routing is restricted to zero-data-retention endpoints, so none of the text sent is retained by the inference provider after the request completes.

You can request further information about the safeguards in place by contacting liebyrinth.support@gmail.com.

6. Data retention

Personal data is retained according to the following criteria:

Account data and project content — retained for as long as the account is active.
Deleted projects — a project you delete is kept in a soft-deleted state for 30 days, during which it can be restored on request; after that period the content is permanently erased.
After an account deletion request — the account is deactivated and all personal data is permanently erased within 30 days of the request.

You can request deletion of your account and all associated data at any time by writing to liebyrinth.support@gmail.com.

7. Minimum age

Bloomwright is intended for users who are at least 14 years old, in accordance with Art. 2-quinquies of Italian Legislative Decree 101/2018. If you are under 14 you may not create an account or use the service. The controller reserves the right to delete accounts where there is reasonable doubt that this requirement is met.

8. Your rights as a data subject

Under Articles 15–22 GDPR, you have the right to:

Access (Art. 15) — obtain confirmation that we process your data, and a copy of it.
Rectification (Art. 16) — correct inaccurate or incomplete data.
Erasure (Art. 17) — request deletion of your data (the "right to be forgotten").
Restriction of processing (Art. 18) — request that processing be restricted in certain cases.
Portability (Art. 20) — receive your data in a structured, commonly used, machine-readable format.
Objection (Art. 21) — object to the processing of your data on legitimate grounds.

To exercise your rights, write to liebyrinth.support@gmail.com. We will respond within 30 days of receiving your request.

You also have the right to lodge a complaint with the supervisory authority: the Garante per la protezione dei dati personali (the Italian data protection authority), Piazza Venezia 11 — 00187 Rome, www.garanteprivacy.it, email: garante@gpdp.it.

9. Cookies and local storage technologies

For detailed information about our use of cookies and localStorage, see our Cookie Policy.

10. Page-view measurement

We count page views on our own servers, so we know how many people reach Bloomwright and which pages they read. For each page view we record the page path, the host of the site that referred you (for example google.com — never the full address), and a random session identifier generated by your browser.

That identifier is kept in your browser’s sessionStorage and is discarded when the browser session ends. It is not a cookie, it is never stored on our servers alongside your account, and it tells us nothing about you beyond the fact that two page views belong to the same visit. If you sign up during a visit, it is sent once more so we can look up the page that visit started on — see Arrival page above — and it is discarded after that lookup.

We do not store your IP address for this purpose — it is used only to rate-limit abusive requests and is then discarded — and this measurement is not shared with any third-party analytics provider.

11. Changes to this notice

The controller reserves the right to update this notice to reflect changes in the law or in the service. In the event of material changes, a notice will be shown inside the application the next time you sign in, and may also be sent to the email address associated with your account. The current version will always be available on this page, with the date it was last updated.