Privacy policy
HERBALISTAI
PRIVACY POLICY
Effective date: 2026 11 03 | Last updated: 2026 11 03
1. Introduction and controller identity
This Privacy Policy explains how personal data is collected, used, stored, disclosed and otherwise processed when you visit the HerbalistAI website, create an account, purchase a subscription or use the Service.
Data controller: Peter V. Rayzman, an individual operating under the business/trading name “RayzmanAI AI systems – development alliance”. The trading name is not a separate legal entity.
Business/contact address: Darinke Radovic 26, Belgrade, Serbia, 11250
Privacy contact: [email protected]
The individual identified above acts as controller for the processing described in this Policy where applicable data-protection law applies. This Policy should be read with the Terms of Use and Cookie Policy.
2. Categories of personal data
2.1 Account and authentication data
• Email address and account identifier.
• Password hash and account status.
• Authentication and security information.
• Account creation and modification timestamps.
• Language and interface preferences, where used.
Passwords should not be stored in reversible/plain-text form. The precise technical controls are subject to verification against the production implementation.
2.2 Subscription and transaction data
• Subscription plan, status, start/end dates and renewal/cancellation information.
• Payment-provider customer, subscription or transaction identifiers.
• Billing events and records needed for reconciliation, support, accounting and legal compliance.
Payment-card information is handled according to the checkout and Payment Provider arrangement. The application is not intended to store complete payment-card details. The Payment Provider’s own privacy notice and contractual role apply to the data it processes.
2.3 Conversations and AI-generated content
• Your messages and AI-generated responses.
• Conversation identifiers, titles, timestamps and associated references/citations.
• Recipe or lifestyle-tip classifications, where generated or stored.
• Verification, safety and operational metadata associated with an answer.
• Technical information required to process and store conversations.
A conversation may reveal information about health, symptoms, medication, lifestyle or personal circumstances. Please avoid submitting personal data that is not necessary.
2.4 Memory information
If memory features are enabled, information derived from conversations may be processed to support continuity between interactions. Memory information may be stored separately or logically distinguished from ordinary conversation history where supported. The Service may provide separate operations for deleting an individual memory, resetting memory and deleting an account; their effects are described in Section 8.
2.5 Technical and security data
• IP address and connection metadata.
• Browser and device information.
• Authentication events and request timestamps.
• Error, performance, security and abuse-prevention records.
The intended production configuration is to avoid recording private conversation content in ordinary application logs. This statement must be checked against actual logging, monitoring and error-reporting settings before publication.
2.6 Uploaded files
If file-upload functionality is available, we may process file name, type and size, upload timestamp, storage location or identifier, file owner, and download/deletion events. At launch, uploaded user documents are not intended to be automatically incorporated into the public HerbalistAI knowledge corpus. Files must be protected by access controls so that one user cannot access another user’s files.
2.7 Support and administrative communications
If you contact us, we may process your contact details, the content of your request and information needed to investigate and respond. We may retain records of material account or administrative actions for security, accountability and dispute resolution.
3. Health-related and other sensitive information
HerbalistAI is an educational wellness service, not a medical provider. Nevertheless, a user may voluntarily disclose information that qualifies as health data or another special category of personal data under applicable law.
We do not require you to disclose your medical history to register. If you submit health-related information, it may be processed to generate a response, maintain conversation history, apply relevant safety rules and provide features you request.
HerbalistAI does not collect or process of health data.
If you do not wish to submit health-related information, you may use the Service without including unnecessary personal medical details in your questions.
4. Purposes and legal bases
The legal basis depends on the specific processing activity, the user’s location and applicable law. The table below describes possible bases; it is not a substitute for the controller’s documented assessment. The operator must confirm the basis actually relied upon before publication and operation.
Purpose Typical data Potential legal basis (confirm)
Create and maintain accounts Email, account ID, authentication data Contract performance or pre-contractual steps
Authenticate and protect accounts Authentication and security records Contract, legitimate interests or legal obligation, as applicable
Provide AI chat and retrieve sources Messages, conversation metadata, source references Contract where applicable; separate basis and Article 9 condition where required
Maintain conversation history and memory Messages, timestamps, memory records Contract or another documented basis; Article 9 condition where relevant
Manage subscriptions Plan, status, transaction identifiers Contract and legal obligations
Respond to support requests Contact details and support messages Contract, legitimate interests or legal obligation
Prevent fraud and abuse Security and technical records Legitimate interests or legal obligation
Maintain service reliability Performance and error metadata Legitimate interests, contract or legal obligation
Accounting and legal compliance Transaction and business records Legal obligation
Handle privacy requests Request and identity-verification records Legal obligation and/or legitimate interests
Where processing is based on consent, you may withdraw consent at any time. Withdrawal does not affect the lawfulness of processing carried out before withdrawal. Where we rely on legitimate interests, we will consider your rights and interests and provide the information required by law.
5. AI providers and external services
To operate the Service, information may be processed by external providers for hosting, database storage, language-model processing, payment processing, email delivery, authentication, monitoring, analytics, consent management or other technical functions.
Provider category Provider / configuration
Hosting and infrastructure [INSERT PRODUCTION PROVIDER AND PROCESSING LOCATIONS]
Language-model routing / inference OpenRouter may be used; list the actual production service, models and relevant upstream providers: [CONFIRM]
Payment and subscription processing [INSERT SELECTED MERCHANT OF RECORD / PAYMENT PROVIDER]
Email delivery and authentication [INSERT PROVIDER, IF USED]
Monitoring and error tracking [INSERT PROVIDERS, IF USED]
Analytics, advertising and consent management [INSERT ACTUAL PROVIDERS AND TECHNOLOGIES]
External scientific information sources PubChem and EFSA OpenFoodTox, where used for supporting hazard lookups
When an LLM provider receives a user message, information necessary to generate the requested response may be transmitted to that provider and, depending on the routing arrangement, to an upstream model provider. The data transmitted, processing locations, retention practices and use for model training depend on the provider, model, account settings and contractual terms.
[OPERATOR ACTION REQUIRED: Confirm the production provider chain, provider roles, data sent, retention, model-training settings, processing locations and applicable data-processing/transfer terms. Do not claim that prompts are never retained or used for training unless the relevant provider terms and settings substantiate that claim.]
External scientific databases may receive search terms, herb names or compound identifiers as part of a lookup. The Service is intended to avoid transmitting unnecessary account identifiers or private conversation details to those databases.
6. Sharing and disclosure
Personal data may be disclosed to service providers acting on our behalf; payment or transaction providers; professional advisers, auditors or insurers where reasonably necessary; competent public authorities where required by law; and other parties where necessary to establish, exercise or defend legal claims.
Where a provider processes personal data on our behalf, appropriate contractual and security arrangements must be in place as required by applicable law. We do not sell personal data as a business model. We do not make private conversations available to other users. Access by authorised personnel is limited to legitimate operational, security and support purposes.
7. International transfers
The operator is based in Serbia. Infrastructure and service providers may process personal data in other countries, including countries outside the European Economic Area. Where applicable law restricts an international transfer, the operator will use a legally recognised transfer mechanism, such as an adequacy decision or appropriate contractual safeguards, and implement any additional measures required by law.
[INSERT ACTUAL COUNTRIES, PROVIDERS AND TRANSFER MECHANISMS AFTER CONFIRMING THE PRODUCTION DATA FLOWS.]
8. Retention and deletion
Personal data is retained only for as long as reasonably necessary for the purposes described in this Policy, subject to legal obligations, security needs and the actual retention periods disclosed below. The periods and technical behaviour in this section must be verified against the production implementation before publication.
8.1 Account information
Account information is retained while the account is active and for any additional period required to handle legal obligations, disputes, security matters or other legitimate retention needs.
8.2 Conversation history
Conversation records are retained while associated with an active account unless deleted by the user or a different retention rule is disclosed in the Service. [CONFIRM whether conversations remain in backups after account deletion and the applicable expiry period.]
8.3 Individual memory deletion
The intended memory-deletion process removes the memory from active retrieval and soft-deletes the corresponding database record. The draft technical policy specifies a 30-day restoration period during which the record may remain recoverable. A soft-deleted memory must not be used to personalise subsequent responses. [VERIFY this behaviour and retention period against production.]
8.4 Reset all memory
Resetting all memory applies the memory-deletion process to memories associated with the account. It does not necessarily delete ordinary conversation history; these are separate data categories.
8.5 Account deletion
The intended process deactivates the account and prevents login during a 30-day grace period. After that period, the account-deletion process is intended to delete the live account record, conversations, messages, memories, files and sessions, subject to information that must lawfully be retained. [CONFIRM whether users can cancel the deletion request during the grace period and ensure the interface and implementation match this description.]
8.6 Backups
The intended backup policy retains encrypted backups for up to 30 days. Data deleted from the live system may remain in a backup until that backup expires. Backups are not designed for selectively restoring one deleted account or record. If a backup is restored, appropriate deletion records or controls should prevent previously deleted data from being unintentionally reintroduced. [VERIFY encryption and actual backup-retention settings.]
8.7 Scientific hazard lookup cache
The intended active-cache retention period for scientific lookup information from PubChem and EFSA OpenFoodTox is 180 days. Older records may be archived for scientific provenance, operational traceability and audit purposes. The cache should contain scientific lookup information rather than private conversation histories or account data. [VERIFY actual fields, archive retention and whether any personal data is present.]
8.8 Legal and accounting records
Transaction, accounting, security and administrative records may be retained for longer where required by law or reasonably necessary to establish, exercise or defend legal claims. Access will be limited to the relevant purpose.
9. Your data-protection rights
Depending on your location and applicable law, you may have the right to request access, rectification, erasure, restriction of processing, data portability, object to certain processing, withdraw consent and lodge a complaint with a competent supervisory authority. These rights are subject to applicable legal conditions and exceptions.
To exercise a right or ask a privacy question, contact [email protected]. We may need to verify your identity before responding. Where the GDPR applies, you may lodge a complaint with the supervisory authority competent for your circumstances. The Serbian Commissioner for Information of Public Importance and Personal Data Protection may also be relevant where Serbian law applies.
10. Cookies and similar technologies
Cookies and similar technologies are described in the separate Cookie Policy. Strictly necessary technologies may be used where permitted to provide security, authentication, session management and core functionality. Non-essential analytics, advertising and similar technologies will be activated only after valid consent where consent is required by applicable law. The actual technologies must be listed in the Cookie Policy and reflected in the consent interface.
11. Security
The operator will implement appropriate technical and organisational measures intended to protect personal data against unauthorised access, loss, alteration and disclosure. Intended controls include HTTPS, access controls, password hashing, restricted access to server-side secrets, database and backup protection, logging configured to avoid unnecessary exposure of private messages, monitoring and recovery procedures. [VERIFY each control actually deployed; remove or amend any control not implemented.]
No system can be guaranteed to be completely secure. Personal-data incidents will be handled in accordance with applicable law.
12. Children
HerbalistAI is intended for adults aged 18 and over. We do not intentionally offer accounts to children. If we become aware that an account has been created by a person below the permitted age, we may take appropriate steps to deactivate the account and delete associated personal data, subject to legal obligations.
13. Changes and contact
We may update this Policy to reflect changes in the Service, providers, legal requirements or processing activities. Where a change materially affects your rights or the law requires notice or consent, we will take the required steps before applying the change. The current version will be published on the HerbalistAI website.
Data controller: Peter V. Rayzman, trading as RayzmanAI AI systems – development alliance
Business/contact address: Darinke Radovic 26, Belgrade, Serbia, 11250
Email: [email protected]
Privacy requests and notices: [email protected]
PRIVACY POLICY
Effective date: 2026 11 03 | Last updated: 2026 11 03
1. Introduction and controller identity
This Privacy Policy explains how personal data is collected, used, stored, disclosed and otherwise processed when you visit the HerbalistAI website, create an account, purchase a subscription or use the Service.
Data controller: Peter V. Rayzman, an individual operating under the business/trading name “RayzmanAI AI systems – development alliance”. The trading name is not a separate legal entity.
Business/contact address: Darinke Radovic 26, Belgrade, Serbia, 11250
Privacy contact: [email protected]
The individual identified above acts as controller for the processing described in this Policy where applicable data-protection law applies. This Policy should be read with the Terms of Use and Cookie Policy.
2. Categories of personal data
2.1 Account and authentication data
• Email address and account identifier.
• Password hash and account status.
• Authentication and security information.
• Account creation and modification timestamps.
• Language and interface preferences, where used.
Passwords should not be stored in reversible/plain-text form. The precise technical controls are subject to verification against the production implementation.
2.2 Subscription and transaction data
• Subscription plan, status, start/end dates and renewal/cancellation information.
• Payment-provider customer, subscription or transaction identifiers.
• Billing events and records needed for reconciliation, support, accounting and legal compliance.
Payment-card information is handled according to the checkout and Payment Provider arrangement. The application is not intended to store complete payment-card details. The Payment Provider’s own privacy notice and contractual role apply to the data it processes.
2.3 Conversations and AI-generated content
• Your messages and AI-generated responses.
• Conversation identifiers, titles, timestamps and associated references/citations.
• Recipe or lifestyle-tip classifications, where generated or stored.
• Verification, safety and operational metadata associated with an answer.
• Technical information required to process and store conversations.
A conversation may reveal information about health, symptoms, medication, lifestyle or personal circumstances. Please avoid submitting personal data that is not necessary.
2.4 Memory information
If memory features are enabled, information derived from conversations may be processed to support continuity between interactions. Memory information may be stored separately or logically distinguished from ordinary conversation history where supported. The Service may provide separate operations for deleting an individual memory, resetting memory and deleting an account; their effects are described in Section 8.
2.5 Technical and security data
• IP address and connection metadata.
• Browser and device information.
• Authentication events and request timestamps.
• Error, performance, security and abuse-prevention records.
The intended production configuration is to avoid recording private conversation content in ordinary application logs. This statement must be checked against actual logging, monitoring and error-reporting settings before publication.
2.6 Uploaded files
If file-upload functionality is available, we may process file name, type and size, upload timestamp, storage location or identifier, file owner, and download/deletion events. At launch, uploaded user documents are not intended to be automatically incorporated into the public HerbalistAI knowledge corpus. Files must be protected by access controls so that one user cannot access another user’s files.
2.7 Support and administrative communications
If you contact us, we may process your contact details, the content of your request and information needed to investigate and respond. We may retain records of material account or administrative actions for security, accountability and dispute resolution.
3. Health-related and other sensitive information
HerbalistAI is an educational wellness service, not a medical provider. Nevertheless, a user may voluntarily disclose information that qualifies as health data or another special category of personal data under applicable law.
We do not require you to disclose your medical history to register. If you submit health-related information, it may be processed to generate a response, maintain conversation history, apply relevant safety rules and provide features you request.
HerbalistAI does not collect or process of health data.
If you do not wish to submit health-related information, you may use the Service without including unnecessary personal medical details in your questions.
4. Purposes and legal bases
The legal basis depends on the specific processing activity, the user’s location and applicable law. The table below describes possible bases; it is not a substitute for the controller’s documented assessment. The operator must confirm the basis actually relied upon before publication and operation.
Purpose Typical data Potential legal basis (confirm)
Create and maintain accounts Email, account ID, authentication data Contract performance or pre-contractual steps
Authenticate and protect accounts Authentication and security records Contract, legitimate interests or legal obligation, as applicable
Provide AI chat and retrieve sources Messages, conversation metadata, source references Contract where applicable; separate basis and Article 9 condition where required
Maintain conversation history and memory Messages, timestamps, memory records Contract or another documented basis; Article 9 condition where relevant
Manage subscriptions Plan, status, transaction identifiers Contract and legal obligations
Respond to support requests Contact details and support messages Contract, legitimate interests or legal obligation
Prevent fraud and abuse Security and technical records Legitimate interests or legal obligation
Maintain service reliability Performance and error metadata Legitimate interests, contract or legal obligation
Accounting and legal compliance Transaction and business records Legal obligation
Handle privacy requests Request and identity-verification records Legal obligation and/or legitimate interests
Where processing is based on consent, you may withdraw consent at any time. Withdrawal does not affect the lawfulness of processing carried out before withdrawal. Where we rely on legitimate interests, we will consider your rights and interests and provide the information required by law.
5. AI providers and external services
To operate the Service, information may be processed by external providers for hosting, database storage, language-model processing, payment processing, email delivery, authentication, monitoring, analytics, consent management or other technical functions.
Provider category Provider / configuration
Hosting and infrastructure [INSERT PRODUCTION PROVIDER AND PROCESSING LOCATIONS]
Language-model routing / inference OpenRouter may be used; list the actual production service, models and relevant upstream providers: [CONFIRM]
Payment and subscription processing [INSERT SELECTED MERCHANT OF RECORD / PAYMENT PROVIDER]
Email delivery and authentication [INSERT PROVIDER, IF USED]
Monitoring and error tracking [INSERT PROVIDERS, IF USED]
Analytics, advertising and consent management [INSERT ACTUAL PROVIDERS AND TECHNOLOGIES]
External scientific information sources PubChem and EFSA OpenFoodTox, where used for supporting hazard lookups
When an LLM provider receives a user message, information necessary to generate the requested response may be transmitted to that provider and, depending on the routing arrangement, to an upstream model provider. The data transmitted, processing locations, retention practices and use for model training depend on the provider, model, account settings and contractual terms.
[OPERATOR ACTION REQUIRED: Confirm the production provider chain, provider roles, data sent, retention, model-training settings, processing locations and applicable data-processing/transfer terms. Do not claim that prompts are never retained or used for training unless the relevant provider terms and settings substantiate that claim.]
External scientific databases may receive search terms, herb names or compound identifiers as part of a lookup. The Service is intended to avoid transmitting unnecessary account identifiers or private conversation details to those databases.
6. Sharing and disclosure
Personal data may be disclosed to service providers acting on our behalf; payment or transaction providers; professional advisers, auditors or insurers where reasonably necessary; competent public authorities where required by law; and other parties where necessary to establish, exercise or defend legal claims.
Where a provider processes personal data on our behalf, appropriate contractual and security arrangements must be in place as required by applicable law. We do not sell personal data as a business model. We do not make private conversations available to other users. Access by authorised personnel is limited to legitimate operational, security and support purposes.
7. International transfers
The operator is based in Serbia. Infrastructure and service providers may process personal data in other countries, including countries outside the European Economic Area. Where applicable law restricts an international transfer, the operator will use a legally recognised transfer mechanism, such as an adequacy decision or appropriate contractual safeguards, and implement any additional measures required by law.
[INSERT ACTUAL COUNTRIES, PROVIDERS AND TRANSFER MECHANISMS AFTER CONFIRMING THE PRODUCTION DATA FLOWS.]
8. Retention and deletion
Personal data is retained only for as long as reasonably necessary for the purposes described in this Policy, subject to legal obligations, security needs and the actual retention periods disclosed below. The periods and technical behaviour in this section must be verified against the production implementation before publication.
8.1 Account information
Account information is retained while the account is active and for any additional period required to handle legal obligations, disputes, security matters or other legitimate retention needs.
8.2 Conversation history
Conversation records are retained while associated with an active account unless deleted by the user or a different retention rule is disclosed in the Service. [CONFIRM whether conversations remain in backups after account deletion and the applicable expiry period.]
8.3 Individual memory deletion
The intended memory-deletion process removes the memory from active retrieval and soft-deletes the corresponding database record. The draft technical policy specifies a 30-day restoration period during which the record may remain recoverable. A soft-deleted memory must not be used to personalise subsequent responses. [VERIFY this behaviour and retention period against production.]
8.4 Reset all memory
Resetting all memory applies the memory-deletion process to memories associated with the account. It does not necessarily delete ordinary conversation history; these are separate data categories.
8.5 Account deletion
The intended process deactivates the account and prevents login during a 30-day grace period. After that period, the account-deletion process is intended to delete the live account record, conversations, messages, memories, files and sessions, subject to information that must lawfully be retained. [CONFIRM whether users can cancel the deletion request during the grace period and ensure the interface and implementation match this description.]
8.6 Backups
The intended backup policy retains encrypted backups for up to 30 days. Data deleted from the live system may remain in a backup until that backup expires. Backups are not designed for selectively restoring one deleted account or record. If a backup is restored, appropriate deletion records or controls should prevent previously deleted data from being unintentionally reintroduced. [VERIFY encryption and actual backup-retention settings.]
8.7 Scientific hazard lookup cache
The intended active-cache retention period for scientific lookup information from PubChem and EFSA OpenFoodTox is 180 days. Older records may be archived for scientific provenance, operational traceability and audit purposes. The cache should contain scientific lookup information rather than private conversation histories or account data. [VERIFY actual fields, archive retention and whether any personal data is present.]
8.8 Legal and accounting records
Transaction, accounting, security and administrative records may be retained for longer where required by law or reasonably necessary to establish, exercise or defend legal claims. Access will be limited to the relevant purpose.
9. Your data-protection rights
Depending on your location and applicable law, you may have the right to request access, rectification, erasure, restriction of processing, data portability, object to certain processing, withdraw consent and lodge a complaint with a competent supervisory authority. These rights are subject to applicable legal conditions and exceptions.
To exercise a right or ask a privacy question, contact [email protected]. We may need to verify your identity before responding. Where the GDPR applies, you may lodge a complaint with the supervisory authority competent for your circumstances. The Serbian Commissioner for Information of Public Importance and Personal Data Protection may also be relevant where Serbian law applies.
10. Cookies and similar technologies
Cookies and similar technologies are described in the separate Cookie Policy. Strictly necessary technologies may be used where permitted to provide security, authentication, session management and core functionality. Non-essential analytics, advertising and similar technologies will be activated only after valid consent where consent is required by applicable law. The actual technologies must be listed in the Cookie Policy and reflected in the consent interface.
11. Security
The operator will implement appropriate technical and organisational measures intended to protect personal data against unauthorised access, loss, alteration and disclosure. Intended controls include HTTPS, access controls, password hashing, restricted access to server-side secrets, database and backup protection, logging configured to avoid unnecessary exposure of private messages, monitoring and recovery procedures. [VERIFY each control actually deployed; remove or amend any control not implemented.]
No system can be guaranteed to be completely secure. Personal-data incidents will be handled in accordance with applicable law.
12. Children
HerbalistAI is intended for adults aged 18 and over. We do not intentionally offer accounts to children. If we become aware that an account has been created by a person below the permitted age, we may take appropriate steps to deactivate the account and delete associated personal data, subject to legal obligations.
13. Changes and contact
We may update this Policy to reflect changes in the Service, providers, legal requirements or processing activities. Where a change materially affects your rights or the law requires notice or consent, we will take the required steps before applying the change. The current version will be published on the HerbalistAI website.
Data controller: Peter V. Rayzman, trading as RayzmanAI AI systems – development alliance
Business/contact address: Darinke Radovic 26, Belgrade, Serbia, 11250
Email: [email protected]
Privacy requests and notices: [email protected]