Privacy Policy

Controller: Sydentia Oy, Hämeenkatu 17 B 57, 33200 Tampere, Finland Contact: privacy@sydentia.com Applies to: the Sparky mobile app (com.sydentia.sparky) and the Sparky backend service at api.sydentia.com. Last updated: 21 August 2026

1. Who we are

Sparky is a private, reflective wellbeing companion. Sydentia Oy ("Sparky", "we", "us") is the data controller for the personal data described here. Because we are established in Finland, this policy is written to meet the EU General Data Protection Regulation (GDPR) and the Finnish Data Protection Act.

Sparky is not a medical, clinical, or therapeutic product and is not a crisis service. See our Terms of Service for the full disclaimer.

2. A note on what Sparky is for

Sparky invites you to write personal reflections about your life and to talk things through with an AI companion. This means you may choose to share sensitive information — how you feel, your health, your relationships, your finances. We treat that content as private and describe below exactly how it is handled, who processes it, and how you stay in control.

3. The data we collect

We collect only what the app needs to work. We do not buy personal data, and we do not sell your personal data to anyone.

3.1 Account and identity

• Email address and display name (from email sign-up or from Google/Apple sign-in).
• Sign-in identifiers from Google or Apple (the provider's subject id and email) when you use social sign-in.
• Password — only if you sign up with email/password. It is stored solely as a salted Argon2id hash; we never store or transmit your plaintext password, and the hash is never shared with any third party.
• Public @handle you choose, used to connect with contacts.
• Session tokens (stored hashed on our servers; stored in the device's secure keystore/keychain on your phone).

3.2 Profile you provide

• Optional profile details: year of birth, sex, marital status, country and place of residence, current life status, language. Some of these are treated as sensitive (see §4).

3.3 Reflection and wellbeing content (sensitive)

• Life entries — the journal/reflection text you write.
• Life-area summaries and assessments — including how satisfied, stressed, confident, safe, or fulfilled you feel (0–100 scales) and related notes. This is wellbeing/mental-state data and is sensitive.
• Personality profile — your Big Five results and history, if you take the questionnaire. This is sensitive.
• Memory items — short summaries and semantic embeddings Sparky keeps so your reflections build on each other. You control what Sparky remembers.

3.4 Conversations

• Your chats with the Sparky AI — the message text you send and Sparky's replies are stored on our servers so your conversation has continuity.
• People chats (human-to-human) — if you use the social features, the messages you send to other members are stored on our servers to deliver them, along with read markers and group membership. See §5 for how AI does and does not touch these.

3.5 Contacts and social graph

• In-app contacts you add by @handle, the private "closeness ring" you place them in, contact requests (including any note you write), and any blocks you set.
• We do not access your device address book or phone contacts, and the app does not request that permission.

3.6 Device and technical

• Push token for your device (to deliver notifications), plus platform and app language. This is a persistent device identifier.
• We do not collect precise location, advertising IDs, or hardware identifiers, and the app contains no third-party analytics, advertising, or crash-reporting SDKs.

3.7 Purchases

• Subscription and in-app-purchase records — store, product, transaction id, and the store's purchase token/receipt, used to verify and maintain your subscription. Card/payment details are handled by Apple or Google, not by us.

3.8 Sponsored content and preferences

• Your ad preferences (whether you allow sponsored suggestions, personalization, and — separately — health- or finance-related suggestions), and which advertisers/suggestions you have hidden.
• Interaction events with sponsored suggestions (shown, tapped, dismissed, hidden, reported). See §6 for how these are kept unlinkable to advertisers.

3.9 Safety

• Crisis-safety signals. To keep you safe, incoming messages are screened for signs of acute distress. We store only the signal metadata (that a signal occurred, its severity, and what the app did) — never the message text — and this is highly sensitive and access-restricted.

3.10 Logs

• Operational and audit logs and an AI-usage cost ledger. By design these hold only references/identifiers and metadata (timestamps, counts, costs) — never the text of your reflections, chats, or messages.

4. Sensitive (special-category) data

Sparky is designed for you to reflect on your wellbeing, so you may provide data that GDPR treats as special-category — in particular data concerning health/mental wellbeing, and revealing details such as sex or marital status. Your Big Five personality profile and your wellbeing assessments are also sensitive by nature. We process this data only to provide the reflection features you choose to use, and our legal basis is your explicit consent (see §7), which you can withdraw at any time by turning off the relevant feature or deleting the content or your account.

5. How AI is used, and what is sent to OpenAI

Sparky's AI replies, summaries, and reflections are generated using OpenAI as our AI processor. Here is exactly how your content is and is not used:

• When you use an AI feature (chat with Sparky, generate a summary or assessment from an entry, run a guided reflection/"lens", or the safety screen), the relevant reflection or message text and minimal profile context are sent to OpenAI to produce the result. Your email, password, and account identity are never sent to OpenAI.
• The models used are OpenAI's gpt-4o-mini (by default) for text and text-embedding-3-small for embeddings.
• People-chat (human-to-human) messages are not sent to OpenAI during normal messaging. They are only processed by AI if a subscriber turns on the optional "Sparky listens" setting for a specific chat. That setting is off by default, requires an active paid subscription, applies only to that one chat, and produces private insights only for the person who turned it on. When it is on, the recent messages in that chat (including other members' messages) may be sent to OpenAI to generate that person's private insight. The disclosure shown before you use people chats explains this.
• We do not use your content to train third-party AI models. OpenAI processes the content to return Sparky's response on our behalf as a data processor.

6. Sponsored content, and why advertisers can't identify you

Sparky may show clearly labelled sponsored suggestions. We never share your identity or your reflection/chat content with advertisers, and no user data is sent to any external advertising network. Ad matching happens on our own servers.

Where any record could relate to an advertiser, we replace your identity with a one-way, keyed pseudonym (an HMAC) that is different for every advertiser, so the same person cannot be linked across advertisers, and the pseudonym cannot be reversed to your account. Any reporting we provide to a partner is aggregated and k-anonymised (small groups are suppressed) and never includes your wellbeing, personality, message, or reflection data. You can turn off sponsored suggestions and personalization, or hide individual advertisers, at any time in Settings.

7. Legal bases (GDPR Article 6 and 9)

• Performance of a contract (Art. 6(1)(b)) — creating and running your account, storing and processing your reflections and messages to deliver the features you use, and managing subscriptions.
• Explicit consent (Art. 6(1)(a) and Art. 9(2)(a)) — processing sensitive wellbeing/health, personality, and profile data; the optional "Sparky listens" AI evaluation of people chats; and personalized sponsored suggestions. You can withdraw consent at any time.
• Legitimate interests (Art. 6(1)(f)) — keeping the service secure, preventing fraud and abuse, and safety screening for signs of acute distress. Where safety processing involves sensitive data, we also rely on protecting your or another person's vital interests (Art. 9(2)(c)).
• Legal obligation (Art. 6(1)(c)) — retaining certain transaction and accounting records.

8. Who processes your data (sub-processors) and international transfers

We use a small set of processors, each limited to a specific purpose:

Where a processor is outside the EU/EEA (currently OpenAI, FCM, Google/Apple), transfers are protected by the European Commission's Standard Contractual Clauses and appropriate safeguards. Our current sub-processor register is available on request. We update this list when a processor changes.

9. How long we keep data, and deletion

• We keep your personal data while your account is active.
• You can delete your account at any time in the app (Settings → Delete account), or by contacting us at privacy@sydentia.com. Deletion is immediate and irreversible.
• When you delete your account, we permanently delete your reflections, summaries, assessments, memories, personality profile, AI-chat history, contacts and social graph, devices/push tokens, ad preferences, and safety signals.
• Some records are retained but de-identified or minimised, because we cannot link them to you afterwards and we need them for legal, accounting, or shared-integrity reasons:
  • Financial/transaction and AI-cost records are kept as monetary aggregates with your identity removed.
  • Messages you sent in group chats are stripped of your content and identity ("tombstoned") so the other participants' conversation remains intact.
  • A security/audit log that does not contain your account identifiers is retained.
• We retain operational logs only for as long as necessary for the purposes described above. (See also our internal note on defining explicit retention limits for long-lived logs.)

10. Your rights

Under the GDPR you have the right to access, rectify, erase, restrict, and object to the processing of your personal data, to data portability, and to withdraw consent at any time (without affecting processing already carried out). Many of these you can exercise directly in the app (edit your profile and reflections, control what Sparky remembers, manage ad preferences, delete your account).

Note: a self-service data-export ("download your data") feature is planned but not yet available in-app; until it ships, you can request a copy of your data by emailing privacy@sydentia.com.

To exercise any right, contact privacy@sydentia.com. You also have the right to lodge a complaint with your supervisory authority; in Finland this is the Office of the Data Protection Ombudsman (Tietosuojavaltuutetun toimisto, tietosuoja.fi).

11. Security

We protect your data with encryption in transit (TLS), strict per-user access controls (row-level security so each account can only reach its own data), salted Argon2id password hashing, and secure token storage on your device. Passwords and reset tokens never leave our own systems. No method of transmission or storage is completely secure, but we work to protect your information and to limit who and what can access it.

12. Children

Sparky is not directed to children. You must be at least 16 years old to use Sparky. We do not knowingly collect personal data from children under this age; if you believe a child has provided us data, contact us and we will delete it.

13. Changes to this policy

We may update this policy as the app evolves. We will post the updated version with a new "Last updated" date and, for material changes, notify you in the app.

14. Contact

Questions or requests: privacy@sydentia.com Sydentia Oy, Hämeenkatu 17 B 57, 33200 Tampere, Finland