MemoFlow Privacy Policy

Version: v1.2 Effective date: October 3, 2026 Operator: Nanjing Junhe E - Commerce Co., Ltd.
Contact: service@memoflow8.com

1. Scope

This Policy explains how MemoFlow processes information when you use its mobile applications, website, APIs, note and sync services, AI features, meeting recording, live interpretation, OCR, membership, AI Credits, and customer support. Third-party stores, identity providers, websites, and services process information under their own policies.

2. Information we process

Depending on the features you choose, we may process:

• account and identity information, including email address, password hash, verification status, login time, Apple or Google identifiers, profile information, language, and settings;
• notes and files, including titles, text, blocks, attachments, exports, versions, and content you select for an AI request;
• voice and interpretation data, including recordings, audio segments, transcripts, speaker labels, source and target languages, and saved session results;
• OCR input and output. Standard OCR runs on the device. Advanced OCR uploads the single image you choose to MemoFlow's servers, where it is recognized in the cloud (normally by a document-recognition model MemoFlow runs itself, or, if we switch providers, by a cloud document-recognition service provider), and returns structured text. Content you later save as a note or use with cloud features may be uploaded at your direction;
• AI activity, including prompts, selected context, responses, citations, model and provider choices, confirmations, tool activity, and operational metadata;
• purchase and entitlement records supplied by Apple or Google or generated by MemoFlow, such as product identifier, transaction status, subscription period, plan, included-usage balance, and refund or revocation state. We do not receive your full payment-card number from those stores;
• device and operational data, including a random app-installation identifier, platform and phone/tablet category, app version and build, operating-system version, device model identifier, app and system language, first/last-seen times, network and request metadata, error codes, performance data, security events, and necessary server logs. We do not use advertising IDs, IMEI, hardware serial numbers, MAC addresses, or device fingerprinting for this purpose; and
• support information you submit, including your reply address, category, message, source page, and request reference.
• launch-notification information you submit, including your email address, language, source page, campaign tag, consent version, confirmation status, and confirmation or unsubscribe timestamps. Confirmation and unsubscribe tokens are stored only as cryptographic hashes.

MemoFlow may keep local notes, recordings, pending uploads, models, and caches on your device to support offline work and reliable upload. Clearing a cache does not necessarily delete notes or pending originals.

3. Why we process information

We process information to provide and synchronize the service; authenticate accounts; complete requested AI, voice, translation, OCR, export, and sharing actions; verify purchases and maintain Credits ledgers; secure and troubleshoot the service; answer support requests; prevent abuse and fraud; comply with law; and improve reliability and user experience where permitted.

If you request a launch notification, we use double opt-in to verify control of the address and send product-launch or essential related information. We do not add that address to unrelated marketing lists. You may unsubscribe through the link provided in an applicable message.

Where applicable law requires consent, we request it before the relevant processing. You can withdraw consent through the available device, app, or account controls, without affecting processing already lawfully completed.

4. AI, voice, and external processing

When you choose a cloud AI, transcription, translation, search, or refinement feature, the content required to perform that request may be sent to the configured service provider. MemoFlow is designed to limit a request to the context you select, but you should not submit secrets or information you are not authorized to process.

For cloud voice features, MemoFlow may durably store accepted audio chunks before sending audio to the configured speech provider. This separates proof that MemoFlow received audio, playback and billing records from proof that the provider processed it. Live Interpretation sends necessary audio, language settings, session identifiers, and operational metadata for realtime source transcription and translation. Its completed realtime text is saved directly; it is not sent through a separate post-session refinement or translation-backfill pass. Audio, text, session metadata, gaps, and usage evidence remain subject to the retention and deletion rules below.

Before MemoFlow sends personal content to a third-party AI service for the first time, the app identifies the data involved and the recipient and asks for your explicit permission. Depending on the feature, the data may include your prompt, necessary conversation history, notes you explicitly select, transcript text, extracted attachment text, a minimized web-search query, model choice, and necessary request metadata. The recipient is DeepSeek for directly provided models, or OpenRouter and the provider of the model you select. If you decline or later withdraw permission in Settings > AI & Privacy, MemoFlow blocks new third-party AI requests while keeping non-AI features available. A material change to the disclosed recipients or data categories requires renewed permission.

Advanced OCR. Standard OCR keeps the image on your device. If you choose Advanced OCR, the app explains this and asks for your confirmation before the first upload (and again if the explanation changes in a material way). The image you select is then sent over an encrypted connection to MemoFlow's servers. Recognition normally runs on a document-recognition model that MemoFlow operates on its own cloud servers (hosted with our infrastructure provider); the model reads the image and returns text and layout structure. No additional large language model summarizes, rewrites or translates the result. MemoFlow keeps the uploaded copy only until the recognition task finishes, fails, is cancelled or times out, and then deletes it; it is not used for anything else. If our own recognition service is unavailable, we may decide to switch new Advanced OCR tasks to a cloud document-parsing service operated by Baidu AI Cloud, which then processes the image on our behalf; in that case our servers delete their copy as soon as the provider has accepted the task, and the provider may retain processing data, and result links it creates may remain valid, for a limited period under its own terms. MemoFlow keeps the structured result for up to 24 hours so you can recover it after a lost connection, and erases it sooner when you finish with it; the job record (identifiers, status, timings and error codes, without the image or the text) is kept for about 30 days for billing reconciliation and support. MemoFlow does not use your OCR images to train models.

AI output, transcription, translation, and OCR may be inaccurate. Review important names, numbers, decisions, and high-impact information. Before recording or transcribing other people, provide notice and obtain all consent required by local law or policy.

5. Service providers and disclosures

MemoFlow uses service providers for infrastructure, storage, email delivery, authentication, app distribution and payment, AI models, embeddings, speech processing, and related operations. These include Tencent Cloud, Cloudflare or compatible object storage, Apple, Google, Resend, OpenRouter and model providers available through it, DeepSeek, Baidu AI Cloud (document recognition for Advanced OCR, only if we switch Advanced OCR to it), and locally deployed or contracted speech and embedding services, depending on the feature and production configuration.

We disclose only the information reasonably needed for the requested function, security, payment verification, support, or legal obligation. We do not sell personal information for money.

6. Purchases and AI Credits

Apple App Store or Google Play processes payment credentials and store billing. MemoFlow receives transaction information needed to validate entitlements, restore purchases, issue Credits, handle refunds or revocations, and prevent fraud. Credits are service units rather than money; their commercial rules are described in the Subscription Terms.

7. Retention and account deletion

We retain information only as long as reasonably needed for the purposes described here, including service delivery, security, fraud prevention, accounting, legal compliance, and dispute handling. Different records require different periods. Advanced OCR upload copies are deleted from MemoFlow's servers when the recognition task ends (or, if a third-party provider is used, once it accepts them); structured results are kept up to 24 hours (or until you finish with them); job records without image or text are kept about 30 days. Local device data remains until you remove it, clear app data, or uninstall the application, subject to device behavior and backups.

You may start account deletion in MemoFlow Settings or use https://memoflow8.com/account-deletion. Deleting an account does not cancel an App Store or Google Play subscription. The current backend immediately disables the account, revokes sessions, releases identity links, and replaces the original email; complete production release also requires a verified task that deletes or de-identifies associated data except where retention is legally necessary. The final published Policy must state the validated schedule.

8. Security

We use measures such as encrypted transport, password hashing, access controls, least privilege, private object storage with short-lived access links, token revocation, request references, validation, rate limiting, and log redaction. No system is perfectly secure. Contact us promptly if you suspect unauthorized access.

9. International processing

The first website and service release targets markets outside Mainland China. Information may be processed in countries or regions where MemoFlow and its providers operate. Advanced OCR images are processed on MemoFlow's own cloud servers, which may be located outside your country or region; if we switch Advanced OCR to Baidu AI Cloud, that provider processes images on infrastructure it operates, which may be located in Mainland China. Before launch in a market, we will confirm required notices, contractual safeguards, consent, or transfer mechanisms.

10. Your choices and rights

Subject to applicable law, you may request access, correction, deletion, restriction, objection, portability, or withdrawal of consent. You can manage some information in the app or email service@memoflow8.com. We may verify account ownership and may retain or decline information where law permits or requires it. You may also complain to the competent data-protection authority.

11. Children

MemoFlow is not directed to children under 14. A person below the legal age to consent in their location may use the service only with required parent or guardian authorization and supervision. Contact us if you believe a child provided information without valid authorization.

12. Changes and contact

We may update this Policy when the product, processing, providers, or law changes. Material changes will receive appropriate notice and a new effective date, and we will request consent where required.

Operator: Nanjing Junhe E - Commerce Co., Ltd.
Registered address: No. 78, Shijiu Hubei Road, Gaochun District, Nanjing, Jiangsu Province, China
Business address: No. 78, Shijiu Hubei Road, Gaochun District, Nanjing, Jiangsu Province, China
Privacy contact: service@memoflow8.com
Privacy Policy: https://memoflow8.com/legal/privacy
Account deletion: https://memoflow8.com/account-deletion