Sprout Privacy Policy

Updated and effective: October 31, 2026

1. Who provides Sprout

Sprout (点滴) uses the support brand perdur and is provided by the individual developer identified on its App Store listing. This policy covers spending and income records, food and water logs, study plans and timers, trends, and optional AI features. Contact us about privacy or support at zzz2327747158@gmail.com.

2. Records on your device

Your salary, transactions, merchants and notes, meals, water, nutrition goals, study tasks and sessions, reports, and profile details such as nickname, sex, height, weight, and age are stored in the app container on your device by default. We do not provide cloud synchronization of your journal. Whether device backups include app data depends on iOS and your settings.

Widgets read the summaries they need through the app's shared App Group. Study topics and timers may appear on the Lock Screen, Dynamic Island, and local notifications when allowed. You can change system permissions or remove widgets.

3. Offline recognition and optional AI

Offline payment screenshot recognition runs on your device. Images are processed when you select them or run a system Shortcut you have configured. Sprout does not monitor payment notifications, read other apps in the background, browse your entire photo library, or poll your clipboard.

AI image recognition sends the image you select after your consent: custom AI sends it directly to your provider; monthly AI sends it through our disclosed regional server to the model provider. Images may include payment amounts, merchants, transaction dates, food, nutrition details, and anything else visible. Crop or obscure unnecessary names, account details, order numbers, addresses, and faces before sending. Results are drafts; a record is created only when you confirm and save it.

AI reports send daily totals for your chosen 7, 30, or 90 days, the preceding equal-length period, spending-category totals, and goals. They exclude source images, merchant and food names, study topics, nickname, and specific height and weight. Summaries can still reveal financial, dietary, and study patterns and are not guaranteed anonymous. The model returns text and structured analysis; charts and PDFs are produced on your device.

Custom AI uses the HTTPS endpoint, model, and API key you enter. Your phone sends data directly to that provider. The recipient and processing location depend on your configuration and the provider's terms. Sprout shows the destination and asks for permission. Changing the recipient requires new consent.

4. Monthly AI service and processing regions

Monthly AI is available only where the subscription and proxy service have been enabled. The app displays both the server and model recipients before you authorize sharing. Language does not change the recipient. Unvalidated services remain unavailable for purchase.

For the international route, global-api.perdur.cn runs on our Alibaba Cloud server in Singapore and forwards model requests to api.z.ai, using Z.ai Standard GLM-5.3-Flash. The planned mainland China route uses api.perdur.cn on Alibaba Cloud in mainland China and Zhipu’s open.bigmodel.cn. The mainland route is not part of this overseas-first launch; its implementation must be validated before opening. Availability is shown in the app; this policy does not mean that both services are open.

With your authorization, the monthly AI server receives the selected image or report aggregates together with Apple-signed transaction information, including transaction and product identifiers, purchase and expiry information, environment, and verification data. It verifies the signature and queries Apple for current subscription status on every AI request, then forwards only the model input to the fixed model provider. The model does not receive your Apple transaction. We do not receive your Apple Account password or payment-card details.

Images, report inputs, and model results pass through server memory for the request; our service does not write them to disk, request-body logs, or backups. We retain a hash of the subscription identifier, its environment and subscription period, and image/report usage counts to enforce the allowance, for at most 90 days. Short-lived hashed network identifiers are used for rate limiting. Restricted recovery backups may retain usage records for up to 14 further days. Hosting and network operators may process connection, error, and security data under their applicable terms.

Z.ai states that personal data is generally processed in Singapore, subject to its applicable policy and service arrangements. Its API data-processing terms describe real-time processing without storage of API content. This is not a promise that all account, network, safety, or operational data has zero retention. Mainland Zhipu's policy describes storage in mainland China and retention for the necessary purpose and legal requirements; it also describes certain anonymized data use for service improvement and training. We do not make a blanket promise of zero retention or no training across every provider or product. Review the applicable API terms and policies before sharing.

Z.ai API privacy policy: https://docs.z.ai/legal-agreement/privacy-policy

Zhipu privacy policy: https://docs.bigmodel.cn/cn/terms/privacy-policy

Zhipu user agreement: https://docs.bigmodel.cn/cn/terms/user-agreement

5. Consent, credentials, and subscriptions

AI consent is tied to the disclosed destination and disclosure version. Monthly AI consent identifies both the proxy server and model recipient. Material changes require renewed consent. You can withdraw consent in user settings or the data-use information screen. New AI uploads then stop and the app returns to offline mode. Existing local records and configuration remain. Withdrawal cannot recall data already transmitted. Contact us or the actual provider about previously transmitted data; do not send API keys or your complete journal.

Custom API keys are stored in the device Keychain and sent to your chosen endpoint for authentication. They are excluded from journal exports, PDFs, and widgets. Keychain items may remain after uninstalling; clear the custom configuration in AI settings to remove its key.

Monthly AI is an optional Apple auto-renewable subscription. Apple handles payment, restoration, renewal, and refunds. The app reads verified transactions and checks eligibility through our server as described above. Complete signed transactions are not persisted, and request and response bodies are not written to business logs. The service does not host your journal or its history; it receives only the image or report aggregates needed for each request.

Operator model credentials stay on the server and are never delivered to the app. The app may cache eligibility and usage information for up to 15 minutes, never beyond subscription expiry; each actual model request still checks current Apple status on the server. This cache is not an API key. Server credentials and recovery backups are restricted to operations.

6. Permissions and tracking

The system photo picker supplies only the images you choose. Study countdowns request local notification permission when needed. This version does not request location, contacts, microphone, camera, or HealthKit access. It has no advertising or behavioral analytics SDK and does not perform advertising tracking. You can photograph a meal with the system Camera before selecting the photo or running screenshot recognition.

7. Retention, export, and deletion

You control the retention and deletion of local records. You can export journal JSON; JSON import is not currently supported. Report PDFs must be shared separately and are excluded from the JSON backup. Deleting a task preserves existing study logs. Uninstalling removes app-container data but does not automatically remove system backups, shared files, Keychain items, or an Apple subscription.

Source images are recognition inputs and are not saved to the journal or automatically added to Photos. This does not mean a model provider never retains inputs. Recipients you choose control files you export or share. Support messages are used to resolve requests, disputes, and necessary compliance matters and are deleted or anonymized when no longer needed, subject to legal retention obligations.

8. Your choices and security

You can view, edit, delete, and export local records, clear custom AI settings, and withdraw AI consent in the app. Contact us to request access, correction, a copy, or deletion of information we process. We may verify identity in proportion to the request and respond within the applicable legal time limit. We do not hold your complete local journal and cannot remotely recover or edit it.

Sprout uses iOS app isolation, HTTPS, and the Keychain. Protect your device and backups and share only data needed for the feature. We will address and notify users of relevant personal-data incidents as required by law.

9. Website and support

Our public policy and help pages do not require a Sprout account or accept journal uploads and contain no advertising or behavioral analytics scripts. Hosting and network providers process IP addresses, browser requests, and necessary error or security information to deliver and protect the pages. Our current website, perdur.cn, is hosted on our Alibaba Cloud server in Singapore and uses HTTPS. Previously published hosted links remain available for compatibility. Emailing support shares your email address, message, and any attachments you choose to provide.

10. Intended use and updates

Sprout is a personal record-keeping tool. Nutrition estimates and goals are general information for adults, not medical diagnosis, treatment, or investment advice. It is not directed at children. Minors should use it with a guardian's guidance and any necessary consent. Contact us if children's data has been handled inappropriately.

We update this policy for material changes and provide in-app notice or seek new consent when required. Chinese and English describe the same service; please contact us if wording is unclear.