1. Scope and privacy contact
This notice concerns visits to the public information website at weslr.de and its www.weslr.de redirect, the public NEWI research conversation, and correspondence sent to its public contact address. WESLR operates as a brand/project under the existing e-qubs business, operated by Chirodea-Rosca Arthur-Robert, Mitterfeld 3, 87719 Mindelheim, Germany. The operator is the controller for the processing described here; public contact details are also in the Impressum. Privacy requests can be sent to [email protected].
The public site now includes Utile free-beta account registration, saved refund cases and a feedback form. There is no file upload, checkout or newsletter subscription. Protected chat/staging systems and the owner-private notebook are separate surfaces; this notice does not announce their availability to customers.
2. Website delivery and local logs
An owner-operated Ubuntu server delivers the static pages through Nginx and Cloudflare Tunnel. Cloudflare receives the visitor's connection at its edge before carrying the request to the origin.
Technical request data is processed for delivery, troubleshooting and security. Local Nginx access logging records the network address visible to the origin, request time and request line, status, response size, referrer where supplied and browser user agent. The origin-visible network address is not necessarily the visitor's original address. Error logging may contain technical request context. Cloudflare can separately process visitor IP addresses and request metadata.
The stated purpose is a functioning, secure website and investigation of technical problems; the relevant basis is Article 6(1)(f) GDPR, based on the interests in secure website delivery and troubleshooting.
3. Local-log retention
The inspected local configuration rotates Nginx logs daily and retains 14 archived rotations in addition to the active file. With successful daily rotation, this ordinarily covers roughly two weeks; it is a rotation count, not a guarantee that every record disappears at an exact age. Empty logs and operational interruptions can affect the elapsed period.
This local setting does not determine Cloudflare retention, mailbox retention, separately retained incident records or backups.
4. Cloudflare and other provider facts
Cloudflare is involved in the site's routing, delivery/security and Email Routing. Public responses also carry Cloudflare network-diagnostic reporting headers. These are distinct from a customer account or a newsletter.
Cloudflare's Data Processing Addendum describes its processor terms and international-transfer arrangements. That document allows processing outside the EU/EEA; this site does not promise EU-only processing. The actual account-specific agreement, enabled log products, retention and transfer arrangement have not yet been confirmed for this notice.
5. Email inquiries
[email protected] uses Cloudflare Email Routing to forward incoming mail to a separate mailbox. The destination is a Gmail mailbox (Google); its address is private. Google's Privacy Policy explains its service processing. The account type, applicable processing arrangements, inquiry-deletion settings and any separate mailbox backups remain unconfirmed; EU-only processing is not promised.
Correspondence can contain the sender's details, message, attachments and technical mail metadata. It is handled to address the requested inquiry, scope a proposed project or deal with an existing relationship. Article 6(1)(b) GDPR can apply to requested pre-contractual steps or contract performance; Article 6(1)(f) can apply to other correspondence, and Article 6(1)(c) where a specific legal obligation applies.
There is currently no verified fixed deletion schedule for received/sent correspondence, trash, archives and mailbox backups. This notice therefore does not promise automatic deletion after six or twenty-four months. You may ask for access, correction or deletion using the route below. Any legally necessary retention must be assessed for the particular record and explained where relevant.
For an initial inquiry, describe the task and an anonymized sample. Do not send passwords, access keys or unnecessary sensitive/confidential files. Sending an inquiry does not itself authorize public disclosure or unrestricted reuse of your materials.
6. Browser storage and analytics
The company information pages include no advertising tracker or third-party analytics script. Utile and NEWI use the distinct storage described below. The separate NEWI research conversation uses the storage described below. Cloudflare may add delivery/security behavior at the edge; its account configuration is separate from the application source. No planned analytics provider or automatic account lifecycle is presented as an active feature here.
The Utile-specific information below applies when you use its account or feedback features; it does not change the separate NEWI processing.
NEWI research conversation
The public NEWI conversation connects to one continuing native research learner. An anonymous browser cookie grants access to that browser's conversation. It is marked Secure, HttpOnly and SameSite=Lax, is scoped to the conversation route, and lasts up to 365 days, refreshed when the session endpoint is opened. Returning in the same browser profile restores access while the cookie remains available; another person using that profile can also see its conversation. Clearing the cookie loses browser access but does not erase server records.
The server stores messages, timestamps, delivery identifiers and results, native learning evidence, confirmed phrase meanings, grounded names, attributed statements and corrections. Speakers are attributed through opaque browser identifiers. Conversation transcripts and attributed conversational claims are scoped to that browser speaker; the single continuing learner can retain acquired knowledge. Statements about external people or projects remain attributed reports rather than independently verified facts. Do not submit personal, confidential or sensitive information.
The tab may temporarily store a message draft, an unacknowledged retry request and up to eight recent question-bound grounding drafts in sessionStorage. The message draft and retry request are removed when acknowledged; grounding drafts can remain in the tab for clarification and reload recovery. The application uses no third-party script or language-model service to write NEWI's replies.
No automatic deletion schedule or self-service deletion function is currently implemented for conversation records, native delivery evidence or learned meanings. Clearing browser storage does not delete those server-side records. Use the privacy-request route below to ask about access, correction or deletion; the handling of retained evidence, learned meanings and provider-held copies must be assessed for the particular request.
Access logging is disabled for the NEWI API location. This does not remove technical error records or change the Cloudflare delivery/provider processing described above.
Utile accounts and saved cases
Utile is a free browser beta operated by the controller named above. Creating an account is optional; the public example calculates in your browser without submitting inputs or saving browser storage. An account requires an opaque username, password and acceptance of the free-beta terms (version and acceptance time are recorded). No account email or real name is required. A one-time recovery code is shown after registration; only its hash is retained. Passwords are salted and hashed with scrypt. Save your code securely; it authorizes a password reset. We cannot restore access from an unverified support request.
For the service you request, the server stores your account identifier, saved case labels, opaque references, amounts, dates, redacted evidence notes, source-policy fields, your attestations, explicit corrections, assumptions and revisions, dependencies and case status. Evidence corrections preserve earlier events until case or account deletion. This processing supports your account and requested calculations (Article 6(1)(b) GDPR). Required account fields are necessary to provide saved cases; public information and the example remain available without them. Do not submit credentials, recovery codes, bank account numbers, third-party names, health data or other sensitive/confidential information.
The Secure, HttpOnly, SameSite=Lax __Host-utile_session cookie provides sign-in for up to eight hours. It is strictly necessary for the account service you request; no advertising cookie is used. Session and OAuth credentials are stored hashed on the server. CSRF protection prevents unrelated sites from submitting authenticated actions. The necessary storage follows the strictly necessary service exception in § 25(2) TDDDG; optional analytics consent is separate. Cases are not kept in browser localStorage. Recovery-code and JSON downloads are stored on your device only when you request them; you control those copies.
Account and case data remain on the operator-managed Ubuntu host until you delete them. Cases are scoped to your account and are not published. The operator administers the host and can technically access stored data; this is not end-to-end encrypted storage. HTTPS protects transport through Cloudflare; Cloudflare still handles delivery metadata and terminates the edge connection. The provider and international-transfer limitations above apply. The browser service does not automatically send case content to a bank, mailbox, LLM or model-training service. A client you separately authorize through MCP can receive data within the granted scopes, as explained below.
Browser sessions expire after eight hours; OAuth access tokens after 15 minutes and refresh tokens after seven days. Expired credentials are purged at startup and hourly. Account security events (action, account identifier and time) are retained for 30 days for troubleshooting and protection against misuse (Article 6(1)(f), legitimate interest in secure operation). Login failures can temporarily lock an account. Utile API access logging is disabled at the origin proxy; this does not remove error logs or Cloudflare processing. Rate limits are applied without storing visitor IPs in the application.
In Account you can change your password, revoke all agent grants, withdraw analytics consent or delete the account, its cases, events, revisions and analytics. Export each case before deletion if needed. Deletion removes application rows; an unavailable host must be recovered before it can process a request. Original source records, downloaded exports and feedback submitted separately are not removed by account deletion. There is no automatic off-device account backup in this beta; host or disk failure can lose data. The historical installation test fixtures were removed before launch. Provider-held copies and any legally necessary incident retention must be handled separately rather than claimed erased automatically.
Optional Utile usage analytics
Analytics is off unless you select it. You can use saved cases without consenting. With consent (Article 6(1)(a) GDPR), the application records HMAC-pseudonymized account/event identifiers, action types, server timestamps, source channel and, when submitted, usefulness ratings or explicit decision-help confirmations. It does not include case text, amounts, email addresses, IP addresses or passwords. These events are stored on the same operator host for 30 days, then purged at startup, hourly and before owner statistics are read. Only the owner can view aggregates; test accounts use a separate store. These figures measure consented events, not all visitors, sales or ChatGPT installations.
Withdraw in Account at any time: this erases your account’s collected analytics and leaves cases usable. Earlier actions are not backfilled if you opt in later. Withdrawal does not affect the lawfulness of earlier processing. Voluntary rating and usefulness confirmations are recorded only when consent is on.
Agent connections
If an operator-registered MCP client is authorized, the consent screen names the client and requested case-read/write scopes. The client can then read or change your account’s data within those scopes. Its provider may process the data under its own notice; examine that before authorizing. The connection is optional and can be revoked in Account; sign-out, password changes and recovery also revoke grants. No actual ChatGPT-host qualification or public directory listing is claimed.
Public feedback and support form
The form receives a category, message (10–2,000 characters), acknowledgement that it contains no secrets and an optional reply email. No account is needed. The operator reviews these messages privately; they are never displayed publicly or sent automatically to an AI service. Use only information necessary to explain your request. A reply address is optional and used only to respond, not to subscribe you to marketing.
Questions and requested service support are processed under Article 6(1)(b) where applicable; other feedback and misuse protection under Article 6(1)(f), for the legitimate interests in handling questions and improving service. Privacy requests are handled under the applicable GDPR obligations (Article 6(1)(c)). Messages and reply addresses in the application expire after 90 days and are purged at startup, hourly and when feedback is read or received. Keep the deletion receipt shown after submission: it lets you delete the stored message immediately without an account. Receipt tokens are stored only as hashes. Separately sent or received email follows the mailbox limitations above, not this application expiry. Rights-request correspondence needed to demonstrate lawful handling may require separate justified retention.
7. Your rights and how to request them
You can request access, rectification, erasure, restriction or portability where the legal conditions apply, object to processing on the applicable grounds, and withdraw consent where processing actually relies on consent. You can complain to a data-protection authority; the Bayerisches Landesamt für Datenschutzaufsicht (BayLDA) provides information for private-sector processing in Bavaria.
Write to [email protected] and describe the request and relevant correspondence or time period. A special form, paid account or reason for an access request is not required. Do not send an identity-document copy by default. If there are reasonable doubts about identity, only proportionate additional information needed to protect the person's data should be requested.
The statutory response period is without undue delay and normally within one month of receipt. Where a lawful extension or refusal applies, the person must still receive the required timely explanation. Access and deletion are assessed separately; data needed for an access response must not be erased in a way that defeats that request. Required retention, provider-held copies and any remaining limitations must be explained rather than reporting an unperformed deletion.
8. Retention information still to be confirmed
The local log rotation and the NEWI browser-cookie lifetime are described above. The cookie lifetime is not a server-data deletion period. Inbox, CDN/security-product, conversation-data and backup deletion periods remain unconfirmed. These limits are disclosed so that unverified template periods are not mistaken for actual deletion behavior. Contact [email protected] for questions about a particular record or request.