Zhixu AI

SUPPORT & INFORMATION

Understand where your data goes.

Updated November 3, 2026 · Zhixu AI free beta

1. On-device content

Conversations, displayable reasoning summaries, drafts, confirmed image attachments and downloaded originals are stored on this device. There are no accounts or cloud sync. Unsent drafts and unconfirmed crops are not automatically uploaded to AI services. Selecting an image or applying a template is not a send action.

2. Sending and generating

When you explicitly send or generate, necessary conversation context, selected confirmed cropped images and image prompts pass through Cloudflare to OpenAI for processing. The app receives responses, available summaries or images. Do not submit sensitive content you do not want those services to process. OpenAI may retain inputs, outputs and metadata under its safety, legal and product data policies. We do not promise zero retention. See OpenAI API data controls.

3. Tasks, integrity and metadata

To verify app/request integrity, prevent replay, reduce abuse and manage free quotas, the service processes App Attest public-key identifiers, proofs/signed receipts, assertion counters, session/request integrity data, task IDs, status, timing, quota records and approximate country/region. Network providers encounter IP addresses and necessary request information; some rate-limit identifiers use keyed hashes of IP addresses. Device integrity is not a user account or an absolute guarantee against abuse. See Apple App Attest.

4. Temporary images and retention

Server image tasks and results are generally targeted for cleanup after about 24 hours; orphaned image objects have a separate seven-day lifecycle rule. Scheduling, fault recovery and platform backup/recovery windows may delay deletion. This is not a promise of exact-time deletion or simultaneous removal of every copy. Authentication keys/receipts, revocation markers and necessary quota/security metadata follow separate retention and are not erased by local clearing. Providers may retain data under their own policies. See Cloudflare object lifecycles and Cloudflare storage and recovery.

5. Optional Firebase usage analytics

Usage analytics is off by default. In supporting versions, only your explicit opt-in enables Firebase/Google Analytics collection of basic usage and feature completion counts. Custom events contain only the chat or image category, without completion identifiers. These events do not include conversations, prompts, images, summaries or API keys. Depending on configuration, the SDK also processes device/app-instance identifiers and app/system information. You can stop future collection in settings; this does not instantly delete previously processed data. See Firebase Analytics collection controls and Firebase privacy and security. We provide no advertising or cross-app advertising tracking feature.

6. Deletion and choices

Delete individual conversations or works, or clear app-local records, drafts, attachments and downloaded originals in settings. The app cannot restore cleared content. Resolve active or unknown-outcome tasks first to preserve recovery information. Local deletion does not erase Photos, shared copies, server quotas or data retained by providers, and does not reset free quotas.

7. This website, processing locations and contact

This support website has no advertising/analytics scripts or third-party fonts and sets no marketing cookies. Cloudflare hosts static pages and may process IP addresses, browser requests and security logs necessary to deliver them. Cloudflare, OpenAI and Google may process data outside your region; we do not promise storage in a particular country or absolute confidentiality. For applicable access, correction or deletion questions, contact d20258002@163.com without sending keys or sensitive content. Requests are handled where identification, verification and applicable requirements permit.