Numen Privacy Policy

Last updated: 2026-10-06

1. Who we are

Numen ("Numen", "we", "us", "our") is a private dream-journaling and reflective "vision"/affirmation app for iOS and Android, previously published under the name Reverie. You can record a dream — typed or spoken — and an AI returns a gentle, reflective "reading." You can also generate a dreamlike AI image of your dream or vision. Numen's readings are reflective and supportive only; they are not medical, clinical, religious, or legal advice, and they describe what a dream "may reflect," never what it "means."

This Privacy Policy explains what data Numen collects, why, who processes it, and the choices and rights you have.

Numen is operated by Bilgehan Coskun, an individual app developer (not a registered company).

If you have any questions, contact us at appfactory.acc@gmail.com.

2. Our privacy approach: local-first and anonymous-first

Numen is built to work with no sign-up. By default:

• Your journal entries and app settings live on your device.
• We only sync your journal to the cloud if you choose to create an account. If you never sign in, your entries are never written to our entries store.
• We do not use third-party advertising, we do not track you across other companies' apps or websites, and we do not sell your personal information.

Some features (AI interpretation, voice transcription, AI image generation) require sending the relevant content to our servers and AI providers to work. Those flows are described below.

3. Data we collect, and why

3.1 Account email — only if you sign in

Numen offers an optional account using Sign in with Apple or Sign in with Google, federated through AWS Cognito. We collect your email address only when you choose to sign in. It is used solely as your account identifier so we can sync your journal across your devices. We support Apple's private-relay email addresses. If you never create an account, we do not collect your email.

3.2 Dream/vision text and tags

When you ask for a reading, your dream/vision text, the app language and the "how to address you" preference (Section 3.6) are sent to our AI providers, reached through OpenRouter — Google (Gemini, for readings and transcription), OpenAI (GPT Image, for images) and Meta (Muse Image, the image fallback) — with fal.ai as a backup route for images when OpenRouter is unavailable. Google (Gemini) produces the reading. If you request an image, a short visual plan is derived from your entry by Google (Gemini) and the resulting image prompt is sent to OpenAI (GPT Image), or to Meta (Muse Image) if OpenAI declines it; if OpenRouter cannot be reached, the same prompt goes to the same two models through fal.ai. These providers process the content only to return the result.

• If you are signed in, these entries are also stored in our database (AWS DynamoDB) so they can sync across your devices.
• If you are anonymous (not signed in), your entries are not written to our entries store.

3.3 Voice / audio — processed transiently

If you speak a dream instead of typing it, the audio is sent to Google (Gemini) through OpenRouter for transcription, with the request flagged so that the provider does not retain it. The audio is not stored on our servers — it is processed transiently and discarded. The resulting transcript is then treated exactly like text you typed (see Section 3.2).

3.4 Generated images

If you generate an AI image of your dream or vision, the image is created by OpenAI (GPT Image) or Meta (Muse Image) as described in Section 3.2 and stored in AWS S3 so the app can display it by URL. Generated images may carry provenance metadata (C2PA content credentials) added by the provider; we keep it as it is. Image files are stored with unguessable, random (UUID) keys. The storage bucket is configured for public-read access by URL, which means anyone who has the exact, unguessable URL can view that image; the URLs are not listed or indexed publicly.

3.5 Anonymous device identifier

To run our free tier fairly, we generate a random device identifier on your device. We use it to key a server-side credit ledger that tracks free-tier usage (for example, your one free lifetime image). This identifier is random and is not tied to your identity, name, or email.

3.6 How you appear — optional

Numen asks — once during onboarding, and at any time under Settings → "How you appear" — how you would like to be depicted when you appear in your own dream art. The choices are Woman, Man, Another identity, and Prefer not to say.

• Answering is entirely optional. "Prefer not to say" is the default, it is pre-selected, and you can continue without changing it. Nothing in the app is withheld or degraded if you leave it unanswered, and you can change or clear your answer at any time in Settings.
• Your answer is stored in your app settings on your device. We do not store it on our servers, it is not part of your synced journal, and it is not attached to your account.
• Every reading and image request sent to our AI providers carries one form of address along with your dream, because most of the app's languages put the reader's gender into the grammar: "woman" or "man" if you picked one of those, and otherwise "neutral" — for "Another identity", for "Prefer not to say", and if you have not answered. "Neutral" only asks for gender-neutral wording and a neutrally described figure in the artwork; it says nothing about your gender. (A "vision" entry also sends "Another identity" when you picked it.)
• It is never used for advertising, profiling, analytics, or any purpose other than the two described above.

3.7 Purchases and subscription status

If you purchase Numen Premium, subscription status is managed through RevenueCat and the Apple App Store / Google Play. We receive your subscription/entitlement state so we can unlock premium features. We never see or store your card or payment details — those are handled by Apple and Google.

3.8 What we do not collect

We do not use advertising identifiers, we do not run cross-app/cross-site tracking, and we do not use Apple's App Tracking Transparency (ATT) tracking. We do not sell your personal information. We do not request access to your contacts, photos library, or location for tracking purposes.

4. How AI processing works

Numen's readings are generated by Google (Gemini), voice transcription by Google (Gemini), and images by OpenAI (GPT Image) with Meta (Muse Image) as the fallback. All of them are reached through OpenRouter, an API gateway that forwards our requests to the provider; fal.ai is used as a backup route for images only when OpenRouter is unavailable. We may change which model is used within this set of providers as quality improves. When you use these features, the relevant content — your dream/vision text, your app language, your spoken audio (for transcription), the optional "how to address you" preference described in Section 3.6, and the prompts derived from them — is sent to and processed by these providers to produce the result that is returned to you. To create a reading or an image, the text you enter (and, for voice entries, a transcript) is sent to our AI providers, who process it only to return the result. Your journal is stored on your device, and in your backup only if you sign in. Our API keys and webhook secrets live only on our servers, never in the app.

When a reading detects that a dream may carry thoughts of self-harm, the app shows crisis resources (a helpline for your region) instead of a reading. This decision is made by the AI model on the text of that one entry; we record only that a care screen was shown and the app language, never the entry or who you are.

The app identifies these providers and asks for your consent in its "How Numen uses AI" disclosure before your first AI request; you can review the disclosure at any time in Settings.

Before interpretation or illustration, your text is screened by an AI moderation step that blocks disallowed content (sexual content involving minors, graphic sexual content, graphic violence, self-harm or harm instructions, and hateful content).

We do not use your content to train our own models, and we route requests through OpenRouter with provider settings that do not permit training on or retaining your inputs. Each provider's processing of content submitted through its API is governed by that provider's applicable API and data-processing terms — OpenRouter's, Google's for the Gemini API, OpenAI's for the Image API, Meta's for Muse Image, and fal.ai's.

5. Legal bases for processing (EEA/UK — GDPR)

Where the GDPR or UK GDPR applies, we rely on the following legal bases:

• Performance of a contract (Art. 6(1)(b)): processing your dream text/tags/audio to deliver the interpretation, transcription, and image features you request; storing and syncing entries for signed-in accounts; processing your subscription.
• Legitimate interests (Art. 6(1)(f)): operating the anonymous credit ledger to provide and protect the free tier and prevent abuse; keeping the service secure; AI moderation to keep the service safe.
• Consent (Art. 6(1)(a)): where you choose to create an account and provide your email by signing in, and where you choose to submit a dream for AI processing. You can withdraw consent by deleting your account and/or by ceasing to use the AI features.

Your rights

Subject to applicable law, you may have the right to access, correct, delete, port, restrict, or object to processing of your personal data, and to withdraw consent. You also have the right to lodge a complaint with your local data protection authority. To exercise these rights, contact appfactory.acc@gmail.com. Note that for anonymous use, much of your data lives only on your device and is not associated with an identifier we can look up; deleting the app removes that local data.

6. California privacy notice (CCPA/CPRA)

We do not sell and do not share (as those terms are defined under the CCPA/CPRA, including for cross-context behavioral advertising) your personal information, and we have not done so in the preceding 12 months. We do not use your data for cross-context behavioral advertising. California residents have the right to know, access, delete, and correct their personal information, and to not be discriminated against for exercising these rights. To make a request, contact appfactory.acc@gmail.com. The categories of personal information we may process are described in Section 3 (identifiers such as email and a random device id; user content such as dream text, tags, audio, and generated images; the optional appearance preference described in Section 3.6, which stays on your device; and commercial information such as subscription status).

7. Subprocessors and third parties

We rely on the following third parties to operate Numen:

8. Data retention

• Anonymous local data (entries and settings on your device): retained on your device until you delete it in-app or delete the app.
• Signed-in journal entries, tags, and settings (DynamoDB): retained while your account exists, and deleted when you delete your account.
• Generated images (S3): retained while your account exists, and deleted when you delete your account.
• Anonymous credit-ledger row: retained to enforce usage limits; the device-id row is tombstoned when you delete your account (its weekly counters are kept for up to 30 days so deletion cannot reset the limits) and then expires.
• Image jobs: the dream text an image request carries is removed as soon as the image is finished, and the job record expires within a day; the derived visual plan (an English inventory of the dream's visible details, no verbatim text) is cached for up to seven days.
• Voice/audio: not stored — processed transiently and discarded.
• "How you appear" preference (§3.6): stored only in your on-device settings; we retain no copy. It is removed when you clear it in Settings or delete the app.
• Account email: retained while your account exists; deleted when you delete your account.

9. Security

We use reputable cloud infrastructure (AWS) and encryption in transit (HTTPS/TLS) between the app and our servers. Sensitive credentials — the AI-provider API keys (OpenRouter, fal.ai) and webhook secrets — are stored only on our servers and are never embedded in the app. Generated-image URLs use unguessable random keys. No method of transmission or storage is 100% secure, but we work to protect your data using industry-standard safeguards.

10. International data transfers

Numen's backend runs on AWS in the European Union (Frankfurt, eu-central-1), and our AI providers — OpenRouter, Google, OpenAI, Meta and fal.ai — process AI requests on their own infrastructure, which may be located in other countries. Depending on where you use Numen from, your data may be transferred to and processed in countries other than your own. Where required, such transfers are made under appropriate safeguards (for example, the European Commission's Standard Contractual Clauses) offered by our providers.

11. Account and data deletion

You are in control of your data:

• Delete your account in-app. Numen includes an in-app account deletion option (Settings → Delete my data). Deleting your account erases your server-side journal entries, settings, generated images, your credit-ledger row, and your AWS Cognito user. It does not cancel a subscription; cancel it in your App Store or Google Play account.
• Delete your account without the app. Use our account-deletion page (linked from the Support page) or email appfactory.acc@gmail.com from the address you signed in with; we delete the same data within 30 days and confirm by email.
• Delete anonymous local data. If you have never signed in, your data lives only on your device. Remove it by deleting the entries in-app or by uninstalling the app.

If you need help, contact appfactory.acc@gmail.com.

12. Children

Numen is not directed to children under 13, and the minimum age to use Numen is 13+. We do not knowingly collect personal information from children under 13. If you believe a child under 13 has provided us personal information, contact appfactory.acc@gmail.com and we will delete it.

13. Changes to this policy

We may update this Privacy Policy from time to time. When we make material changes, we will update the "Last updated" date above and, where appropriate, provide additional notice in the app. Your continued use of Numen after an update means you accept the revised policy.

14. Contact

Questions, requests, or concerns:

• Email: appfactory.acc@gmail.com
• Operator: Bilgehan Coskun (individual developer)
• Governing law / jurisdiction: the Republic of Türkiye