Sfuma AI Cam Privacy Policy

Effective Date: March 1, 2024
Last Updated: 2026-08-06

This translation is provided for convenience only. The Chinese version shall prevail in the event of any discrepancy.

Notice:

This update adds disclosures about TrueDepth, Face Data, and cloud AI composition, including local-only handling of TrueDepth outputs, the ordinary two-dimensional images transmitted for cloud inference, the processing purpose and model providers, storage and deletion rules, and how to withdraw consent.

1. Scope of This Policy

This Policy applies to the Sfuma AI Cam app, web client, mini program, and any new product forms developed in the future (collectively, the "Product"). Where a specific feature has its own privacy policy, that policy shall prevail for that feature.

2. How We Collect and Use Your Personal Information

(1) Information Required for Core Features

  • Account registration: Mobile phone number and verification code to complete registration, for service notifications and identity verification
  • Image processing services: The images/videos you upload and editing parameters, used for core features such as AI effect generation and image quality enhancement
  • Device information: Device model, operating system, and unique device identifiers (Android ID, IDFA, etc.) to ensure service stability

(2) Information for Optional Features

  • Location: Used for the photo location watermark (city/landmark name) and scene template recommendations; you can turn it off anytime in system settings. We use coarse, kilometer-level positioning, fetched once when you open the camera screen and stopped as soon as a place name is resolved. Only the city/place-name text is sent to our servers (never raw GPS coordinates); it is used solely to generate the current recommendation and is not retained on our servers. The most recent place name is kept on your device for offline watermarks and is deleted when you uninstall the app. The app does not write GPS coordinates into your photos; when editing album photos that already contain location metadata, the original metadata is left unchanged
  • Photo library access: Used to import local images and save results (iOS/Android)
  • Camera access: Used for real-time capture and AR features

(3) TrueDepth, Face Data, and Cloud AI Composition

When you use the front camera, portrait capture, composition assistance, or related features, the Product may use Apple TrueDepth camera and Vision system capabilities to process, solely on your device, the position of a face in the frame, depth or disparity information, portrait-effects mattes, and subject-segmentation information. This processing supports live preview, subject positioning, composition and pose guidance, portrait blur, and foreground/background separation.

TrueDepth depth information and face-analysis results are never uploaded or stored under any circumstances. Cloud features transmit only an ordinary two-dimensional photo frame, which may contain a facial image, and never transmit depth information, face-mesh data, or portrait mattes.

When you actively use AI Composition, Smart Crop, pose recommendations, or another AI feature clearly identified as cloud-processed, the Product applies necessary compression and Base64 encoding to the current ordinary two-dimensional camera frame or a photo you select, transmits it over an encrypted HTTPS connection to the backend operated by Shanghai Qianying Future Artificial Intelligence Technology Co., Ltd. (上海浅影未来人工智能科技有限公司), and then sends it to the actual model-inference provider: Beijing Volcano Engine Technology Co., Ltd. (Volcano Engine) or Alibaba Cloud Computing Co., Ltd. (Alibaba Cloud). The image is used only to analyze the subject, person position, pose, and frame structure in order to estimate or recommend a better AI composition, crop region, and shooting guidance.

The ordinary two-dimensional photo and its Base64 content exist only briefly in request memory as necessary to complete one real-time inference request. Neither we nor the model provider writes them to a database, object storage, file system, business logs, or backups, and they are released after the inference response, request failure, or timeout. Because no persistent image copy is created, you do not need to submit a separate deletion request for these real-time AI composition input images.

This information is not used to identify or verify you and is not used for Face ID, account authentication, advertising, marketing, profiling, behavioral tracking, or model training. Before first use, we provide a separate notice describing the purpose, method, and recipients and obtain your express consent. You may decline or stop using the feature at any time or revoke camera or photo permission in system settings; after you stop or revoke permission, we will not initiate new image uploads.

(4) Information Used to Improve Our Services

  • Usage logs: Frequency of feature usage, page views, click heatmaps, and similar data, used to optimize product interaction design
  • Crash information: Stack traces collected automatically when the app exits abnormally, used to quickly locate and fix issues

3. Data Sharing and Protection

We share information only in the following circumstances:

  • With your explicit authorization
  • In cooperation with judicial authorities for lawful investigations
  • When entrusting partners to perform necessary data processing (under strict confidentiality agreements)

4. Data Storage and Security

  • Retention period: Except for real-time AI composition input images that do not create a persistent copy as stated in “TrueDepth, Face Data, and Cloud AI Composition,” other original images that must be temporarily retained are kept for no more than 72 hours and automatically deleted after processing
  • Encryption: Stored with AES-256 encryption; transmitted using SSL/TLS 1.2 or above
  • Security certification: ISO 27001 information security management system certified

5. Protection of Minors

If you are under 14 years of age, please use the Product under the guidance of your guardian. We do not knowingly collect personal information from children and will delete any such information immediately upon discovery.

6. Third-Party Services

SDKPurposePrivacy Policy
Alibaba Cloud OSSImage file storagehttps://terms.aliyun.com
Volcano Engine model inferenceCloud AI composition inference from ordinary two-dimensional photoshttps://www.volcengine.com/docs/6256/64902
Alibaba Cloud model inferenceCloud AI composition inference from ordinary two-dimensional photoshttps://help.aliyun.com/zh/model-studio/privacy-notice
WeChat Sharing SDKSocial sharing of contenthttps://weixin.qq.com/cgibin/readtemplate?lang=zh_CN&t=weixin_agreement

7. Updates to This Policy

Material changes will be announced through in-app push notifications. Continued use of the Product constitutes acceptance of the updated Policy.

8. Contact Us

Personal Information Protection Officer: Li Anquan

Phone: +86 21-68781323 (weekdays 9:00–18:00)

Complaints email: privacy@qianying.ai

We typically respond within 15 business days.

Shanghai Qianying Future Artificial Intelligence Technology Co., Ltd. (上海浅影未来人工智能科技有限公司)

Address: No. 188 Pudong South Road, Pudong New Area, Shanghai

Contact: support@qianying.ai

China Internet Reporting Center: https://www.12377.cn