Privacy policy

1. Who we are and what this policy covers

MB Sverto, registered at Naugarduko g. 68-3, Vilnius, Lithuania, operates Sverto’s hosted service and is the controller of the personal information we process to provide that service. Contact us at privacy@sverto.com with questions or requests concerning your information.

This policy covers the Sverto website, hosted application and mobile apps when used with our hosted service. Sverto helps you record and understand your financial activity, investments and net worth.

Adults only

Sverto is intended for people aged 18 and over and is not directed at children. If you believe someone under 18 has provided personal information to our hosted service, contact us using the address above so we can investigate and arrange appropriate deletion.

Self-hosted installations

If you connect to a self-hosted server, the operator of that server controls the financial records and other information sent to it. Records sent only to that server are not received by MB Sverto. The server operator’s configuration and policies determine its hosting, authentication, AI processing and retention.

Some mobile SDKs can still communicate with their providers independently of the selected server. In particular, if you enable Android’s optional ML Kit category suggestions, the SDK can send technical and usage telemetry to Google. Contacting our support or visiting our website also involves processing separate from your self-hosted server.

2. Information we process

• Account information: account identifiers, email address, profile name when supplied, authentication and session information, and preferences such as your reference currency.
• Financial records: accounts, balances, transactions, dates, descriptions, categories, holdings, purchase lots, prices, fees, financial institution details and account identifiers you enter or import, such as IBANs or account numbers.
• Trading 212 connection: if you choose to connect your brokerage account, the API credentials, account details and investment activity needed to import its records. If you choose to save credentials for continued access, we retain them in the configured secret store until the connection or account is deleted.
• Uploads and AI interactions: receipts, images, documents, filenames, Myra messages, responses, tool results and your decisions about proposed changes.
• Search requests: search terms sent to our service to locate transactions, assets or other information. Searches made in Myra can also form part of your saved conversation.
• Technical and usage information: request times and routes, response status, errors, service performance, account/resource identifiers included in diagnostics, and information processed by authentication and device SDKs.
• Authentication security information: Clerk processes IP addresses and device/session information and can derive approximate city/country information from the IP address. This does not require GPS access.
• Optional Android ML Kit telemetry: after you enable category suggestions, app and device information, installation and other identifiers, feature-use events, performance metrics and errors relating to that feature.
• Communications: information you choose to send when contacting us about support, privacy or other requests.

You choose which financial records and attachments to provide and whether to use optional features. Some account and technical information is needed to operate and protect the hosted service.

3. Why we use information

We process information to authenticate users; store, synchronise and display their records; calculate balances and investment performance; search information; operate requested imports, connections and AI features; prevent abuse; diagnose failures; and respond to requests.

Where the GDPR applies, the lawful bases for these activities are:

• Providing the requested service: processing necessary to perform our agreement with you, including account access, financial tracking and features you request.
• Legitimate interests: keeping the service secure and reliable, preventing misuse, investigating problems and handling communications, subject to your rights and reasonable expectations.
• Legal obligations: processing required to meet an applicable legal requirement.
• Consent: where required for a particular optional activity. We explain that activity when requesting consent; you can withdraw it without affecting earlier lawful processing.

Advertising and sale of data

Sverto does not display third-party advertising. We do not sell or rent your personal information, share it with advertisers or data brokers, or use your financial records to build advertising profiles.

4. Providers and other recipients

Information is processed by providers used to deliver Sverto:

Railway

Application hosting, databases, caches and file storage, including account records, financial data, uploads and operational logs.

Axiom

Operational logs, traces, errors and performance information used to diagnose problems and maintain the service. Diagnostic records can include account or resource identifiers.

Clerk

Hosted authentication, account identity, sessions and sign-in security information.

Google Gemini API

Cloud AI processing for Myra, receipt/document processing and semantic search, including relevant text, attachments, financial context and generated output.

Google ML Kit

Optional on-device Android category suggestions and the SDK’s technical and usage telemetry after you enable the feature.

Trading 212

Optional retrieval of investment account history using your API credentials, including orders, dividends and cash movements. Your broker continues to process its own records under your relationship with it.

Google Workspace

Email delivery and storage for our support/privacy address, including message content, attachments, sender/recipient details and delivery metadata.

We may also disclose information when required by law or necessary to protect legal rights, investigate abuse or respond to a valid authority request. Any disclosure must have an applicable legal basis.

These providers process information to deliver the services described above. Some also use specific information for their own purposes, such as account administration, security or SDK telemetry, as explained in their privacy notices.

5. Myra and on-device suggestions

Myra sends your request and the relevant context needed to answer it to the Gemini API. Context can include conversation history, attachments and financial records retrieved by Myra’s tools. Your conversations and proposed actions are stored so you can return to them.

Semantic search also uses the Gemini API to turn text, such as transaction descriptions and search queries, into numerical representations. This can happen when descriptions are saved or updated, separately from opening a Myra conversation.

The hosted service uses the Gemini API under paid-service data terms. Google states that paid-service prompts and responses are not used to improve its products. Separate safety processing still applies: its published abuse-monitoring policy describes retaining prompts, context and outputs for 55 days, with restricted human review for suspected misuse. See Gemini API terms and abuse monitoring.

Android category suggestions are off by default and can be enabled in Settings → Category suggestions. The choice is stored on this installation and is excluded from backup restoration. You can continue choosing categories manually without enabling it. Suggestions run on the device using ML Kit. This local feature does not itself send its input or generated output to Google servers, but the SDK sends technical and usage metrics for performance, debugging, maintenance, API improvement and abuse prevention. This is governed by ML Kit’s terms and Google’s Privacy Policy, separately from paid Gemini processing. Saving a financial record to your hosted account is a separate server operation. Turning suggestions off stops new feature requests and cancels pending suggestion work. It does not delete telemetry Google already received or guarantee cancellation of metrics already queued by the SDK.

6. Where information is processed

Our primary Railway application services, PostgreSQL database, Redis service and uploaded-file bucket are hosted in EU West (Amsterdam). Our Axiom logs and traces are stored in US East, so some information is processed outside the EU.

Authentication, AI and other providers can process information in additional countries under their service arrangements. Self-hosted processing locations depend on the server operator’s choices.

International transfer arrangements depend on the provider, service and destination. Axiom’s published data-processing terms include EU standard contractual clauses. Clerk’s and Google’s published terms describe the EU–US Data Privacy Framework where applicable and standard contractual clauses for relevant transfers. These mechanisms address international transfers; they do not mean processing stays within the EU. You can request information about the safeguards relevant to your data at privacy@sverto.com.

Provider terms: Axiom DPA, Clerk DPA, Google processor terms.

7. Retention and deletion

Different types of information have different lifetimes. Deleting a record from the active service does not necessarily erase operational logs, backups or provider safety records at the same moment.

Hosted account and financial records

We retain your account and financial records while your account is active. Account deletion removes your profile and associated financial and conversation records from the active application database. The separate retention periods below apply to operational records and provider copies.

Conversations, uploads and exports

Conversations can be deleted. Deleting a conversation does not itself delete its separately stored files. Account deletion removes the associated database records and requests removal of stored uploads, thumbnails and exports. Storage failures can prevent file removal.

Clerk identity and connector secrets

Account deletion requests removal of your authentication identity and saved connection credentials during the deletion request. Failed external deletions are not automatically retried. You can contact us at privacy@sverto.com to ask us to investigate incomplete deletion.

Background job records

Background processing records can include transaction descriptions or correction text. They do not have an automatic expiry and are not automatically removed when a task finishes or an account is deleted. They remain until separately deleted. You can ask us to include these records in a deletion request using the contact details below.

Axiom logs and traces

Logs and traces are retained for 30 days in our Axiom dataset.

Railway logs

Our hosting plan provides access to the most recent 7 days of logs. This is an access window, not a guaranteed deletion deadline: Railway may retain older logs under its own service arrangements.

Redis operational data

AI usage counters and cached per-user limits expire within 32 days after the relevant update. They can remain after account deletion until their expiry.

Railway volume backups

Backup copies can remain after information is deleted from the active service. Their retention depends on the applicable backup settings and the provider’s recovery arrangements. Active account deletion does not immediately erase these copies, and we do not promise a fixed erasure deadline for them.

Uploaded-file bucket

Files have no automatic age-based expiry. Account deletion requests removal of uploads, thumbnails and exports, with repeated cleanup as described above.

Gemini safety records

Google’s published abuse-monitoring policy specifies 55 days for prompts, contextual information and outputs. These records are separate from conversations stored by Sverto.

Support and privacy correspondence

Ordinary support correspondence is deleted 12 months after resolution. Minimal records needed to demonstrate handling of a privacy request are retained for up to 3 years after closure. Unnecessary identity documents and attachments are removed earlier. Information needed for an actual dispute or applicable legal obligation is retained for that need, then deleted.

8. Security and information on your device

The hosted service uses HTTPS for public app/API connections and authenticated access to user records. Provider credentials are handled through server-side secret storage. File access uses authenticated operations or time-limited links.

The apps can keep local preferences, session information and cached records to support their operation, including offline access where available. Information retained only on your device is distinct from information sent to our service. A self-hosted server’s transport security and access controls depend on its configuration.

No system can guarantee absolute security. Contact us if you suspect unauthorised access to your information.

9. Your choices and rights

Depending on applicable law, you may have rights to access, correct, delete or receive a portable copy of your information; restrict or object to processing; and withdraw consent where processing relies on it. Some rights have legal exceptions.

Contact privacy@sverto.com to exercise a right or raise a concern. You can complain to Lithuania’s State Data Protection Inspectorate (VDAI), or to the data-protection authority in the country where you live, work or believe an infringement occurred.

For records held exclusively by a self-hosted installation, contact that installation’s operator. We can address information MB Sverto itself processes but cannot administer a server we do not operate.

10. Updates and contact

We will update this policy when our processing practices change and update the date at the top of this page. Material changes will be communicated as required by applicable law.