Effective Date: October 14, 2026
Last Updated: October 14, 2026
This Privacy Policy explains how the developer, Max Chao (趙士翔) ("we," "us," or "our"), handles data in connection with the Gestura mobile application ("Gestura" or the "App"). We respect your privacy and aim to describe the App's data practices clearly and transparently.
This Policy applies only to the App. Ollama servers, custom API endpoints, Hugging Face, Google Play, and other third-party services that you choose to access may be governed by their own terms and privacy policies.
Except as expressly described in this Policy, the App does not automatically collect or track information that directly identifies you during normal use, such as your name, email address, telephone number, or precise location.
When you use the App's offline or local features, including text processing, on-device AI inference, image generation, H3-related generation of videos with audio, music generation, and music transcription, the relevant computation is performed on your mobile device. The H3-related feature can process supported text, image, video, and audio inputs locally.
Text, prompts, reference images, source video and audio, model inputs, AI responses, and generated images, videos, audio tracks, music, and music-transcription results are not transmitted to us, MiniMax, Hugging Face, or another external server merely because you use a local feature. Local H3 generation, music generation, and music transcription do not use hosted-inference services.
Images, videos, and audio that you actively select or import for a local feature are used for the on-device processing you request; selecting or importing a file is not consent to upload it.
Local processing may use device memory, caches, temporary files, and local storage. We do not have remote access to content retained only on your device. The separate in-app reporting feature sends only text you submit and the device information listed in Section 1.4, not your local media or its thumbnails. Connected-mode processing, when separately selected, is described in Section 1.3.
Generated videos are saved within the App by default and are not automatically saved to your photo library. When you press the download or export control, a video is written to your photo library or Android-managed media storage. Images that you choose to export and videos that you download outside the App may remain after the App is uninstalled; you control their subsequent retention and deletion.
If you separately enable a photo-library synchronization or backup service on your device, that service may process exported media under its own settings and policy. Such processing is not a cloud-sync function provided by Gestura.
When you actively connect the App to an Ollama server or another custom API endpoint, text, prompts, images, reference images, or other related input that you provide may be transmitted directly from your device to the server you selected in order to generate text, images, or other AI responses.
Unless the App expressly states otherwise, we do not intercept, review, or store content transmitted through your connection to a custom server. However, the server operator or its infrastructure provider may process your input, IP address, request logs, or other network metadata in accordance with its configuration and policies.
You are responsible for evaluating the privacy, security, and data-retention practices of any server, model, or API endpoint you choose. Do not transmit sensitive or confidential information to a server you do not trust.
This connected-mode description does not mean that the H3 on-device video feature or the local music-generation and music-transcription features send your inputs to a custom endpoint; these features perform processing locally.
The App provides a text-based "Report" or "Feedback" feature. Only when you actively press the submit control are the report text and accompanying device and system information sent to the developer database we use. The App does not automatically upload generation records, music-transcription records, or diagnostic data.
An in-app report contains only:
Report content: text you choose to submit, such as feedback, a problem description, or AI text and prompts that you deliberately include. The feature does not accept image, video, audio, or thumbnail attachments.
Device and system information: App version, device model, Android version, and system language setting, sent only with a report you submit.
We use this data solely to diagnose problems, correct errors, improve the stability of AI, image-generation, video-generation, music-generation, and music-transcription features, improve language support, respond to your report, and address reported misuse or safety issues. We do not sell this data or use it for advertising, marketing, cross-service tracking, or advertising profiles.
Where necessary to operate the report database, maintain service security, or comply with legal obligations, the data may be processed by hosting or infrastructure providers acting on our behalf, or disclosed to a competent authority when legally required. Such processing is limited to the necessary purpose.
Before permitting the relevant H3 model download, the App asks whether you are in a restricted territory and uses your answer to decide whether to make the download available. The answer is processed only on your mobile device and is not transmitted to our database, Hugging Face, MiniMax, or another third party as part of the declaration. It is not part of the accompanying diagnostic information in an in-app report. If you deliberately type location information into your report text, that submitted text is processed under Section 1.4.
This local declaration is separate from the IP address and network information that a model host may receive when your device connects to download a model, as described in Section 4.
Local prompts, source images, video and audio, generated images, videos, audio tracks, music, and music-transcription results, model files, caches, and other on-device data may remain on your device depending on the App feature used, your actions, and your Android settings. You can manage or delete this data using available App controls, Android settings, or file-management tools. Clearing App data or uninstalling the App may remove data in the App's private storage.
Images, videos, audio, music-transcription results, or other files that you separately save, share, or export outside the App must be deleted by you from the location where they are stored. Deleting an in-App copy does not necessarily delete an exported copy.
The text reports and accompanying device and system information received through the in-app reporting feature are stored securely and automatically and permanently deleted within six (6) months after receipt. This retention period does not apply to media or other data kept only on your device.
We may retain data for a longer period only where required by applicable law or where reasonably necessary to address an ongoing security incident, legal request, or dispute. The data will be deleted when that purpose ends.
Depending on the data-protection law that applies where you live, you may have the right to request access to, correction of, deletion of, restriction of, or objection to the processing of data you submitted. Where applicable, you may also withdraw consent or lodge a complaint with a competent supervisory authority.
To exercise a right, email zawshishawn@gmail.com and provide, where possible, the approximate date and time of the report, device model, App version, report identifier, or other information that can help us locate the relevant record. We will request only the information reasonably necessary to verify and process your request and will respond as required by applicable law.
Because the App may not require an account or persistent identifier, we may be unable to complete a particular access, correction, or deletion request if we cannot reasonably identify the record associated with your request.
The App does not integrate third-party advertising SDKs, cross-service tracking software, or analytics tools used to monitor user behavior or create advertising profiles, such as Google Analytics or Firebase Analytics.
The App's features, including on-device video generation with audio, are free and are not unlocked by a payment, subscription, or generation-credit purchase. If the App offers voluntary support and you choose to make a support payment through Google Play, Google processes that transaction and may handle payment, Google Account, device, purchase, and transaction information under its own terms and privacy policies.
For a voluntary support transaction, the App may receive limited product and purchase-status information needed to confirm or handle the transaction, not to activate paid features or subscription entitlements. We do not directly receive your full payment-card details. A support payment does not change feature access.
The App may use the Hugging Face API to retrieve public model information, such as model names, descriptions, metadata, and download links. When downloading models or related components for text, image, video or music generation, or music transcription, your device may connect directly to Hugging Face or the relevant model-file host. The H3-related feature downloads developer-provided quantized models and related components from Hugging Face through the App.
Hugging Face and relevant model-file hosts may process your IP address, request logs, device information, or network metadata under their own privacy policies. A download request therefore involves third-party network processing even though generation and transcription are local. Downloaded model files are stored on your device until you delete them through the App, Android settings, or a file-management tool.
Browsing a model list or downloading a model does not cause the App to upload local prompts, reference images, source video or audio, generated images, videos, audio tracks, music, music-transcription results, or thumbnails to Hugging Face or MiniMax. It also does not cause the local region-declaration answer to be transmitted. If you separately use a third-party hosted-inference service, its own privacy policy governs that separate processing; the on-device H3, local music-generation, and music-transcription features are not such services.
https://huggingface.co/privacy
We use reasonable technical and organizational measures to reduce the risk of unauthorized access, loss, misuse, alteration, or disclosure of data received through the reporting feature. However, no method of electronic transmission or storage can be guaranteed to be completely secure.
The security of on-device data also depends on your device, operating system, screen-lock configuration, backup settings, and any custom server you select. We cannot guarantee the security of third-party servers that you control or choose to use.
The App is not directed primarily to children under 13. If a jurisdiction requires a higher minimum age, such as 16, the locally applicable age requirement applies. We do not knowingly collect personal information from children.
If you are a parent or guardian and believe that a child submitted data to us through the reporting feature without appropriate consent, contact us promptly at zawshishawn@gmail.com. Where we can reasonably identify the relevant record, we will take appropriate steps to delete it promptly.
We may update this Privacy Policy from time to time to reflect changes in features, technology, law, or our operations. The revised version will be posted on this page with an updated "Effective Date." If a change materially affects your rights or how data is processed, we may provide an additional notice in the App where appropriate.
We encourage you to review this Privacy Policy periodically.
For questions or suggestions about this Privacy Policy, or to exercise your data rights, contact us at:
Developer: Max Chao (趙士翔)
Email: zawshishawn@gmail.com
生效日期：2026 年 10 月 14 日
最後更新日期：2026 年 10 月 14 日
本隱私政策說明開發者趙士翔（Max Chao）（以下稱「我們」）如何處理您使用行動應用程式 Gestura（以下稱「本 App」）時涉及的資料。我們重視您的隱私，並致力於以清楚、透明的方式說明本 App 的資料處理方式。
本政策僅適用於本 App。您自行選擇連線的 Ollama 伺服器、自訂 API 端點、Hugging Face、Google Play 或其他第三方服務，可能另受其各自的條款與隱私政策規範。
除本政策明確說明的情形外，在正常使用情況下，本 App 不會由我們自動收集或追蹤可直接識別您身分的資訊，例如姓名、電子郵件地址、電話號碼或精確位置。
當您使用本 App 的離線或本地功能，包括文字處理、裝置端 AI 推論、圖片生成、H3 相關有聲影片生成、音樂生成及轉譜時，相關運算會在您的手機上執行。H3 相關功能可在本地處理支援的文字、圖片、影片及音訊輸入。
文字、提示詞、參考圖片、來源影片與音訊、模型輸入、AI 回應，以及生成圖片、影片、音軌、音樂與轉譜結果，不會因使用本地功能而傳送給我們、MiniMax、Hugging Face 或其他外部伺服器。H3 本地生成、本地音樂生成與轉譜均不使用雲端託管推論服務。
您為使用本地功能而主動選擇或匯入的圖片、影片與音訊，係用於您要求的裝置端處理；選擇或匯入檔案，不代表同意上傳。
本地處理可能使用裝置記憶體、快取、暫存檔及本地儲存空間。我們無法遠端存取僅保留在您裝置上的內容。獨立的 App 內回報功能只會傳送您提交的文字及第 1.4 節所列裝置資訊，不會傳送本地媒體或其縮圖。您另行選擇連線模式時的資料處理，依第 1.3 節說明。
生成影片預設保存在 App 內，不會自動存入相簿。當您按下下載或匯出操作，影片才會寫入相簿或由 Android 管理的媒體儲存空間。您選擇匯出的圖片及下載至 App 外部的影片，在移除本 App 後仍可能保留；後續保留與刪除由您控制。
若您另行在裝置上啟用相簿同步或備份服務，該服務可能依其設定與政策處理已匯出的媒體。此類處理不是 Gestura 提供的雲端同步功能。
當您主動將本 App 連線至 Ollama 伺服器或其他自訂 API 端點時，您提供的文字、提示詞、圖像、參考圖像或其他相關輸入內容，可能會直接從您的裝置傳送至您所指定的伺服器，以產生文字、圖片或其他 AI 回應。
除非本 App 介面另有明確說明，我們不會在該連線過程中攔截、檢視或儲存您傳送至自訂伺服器的內容。然而，該伺服器的營運者或基礎設施提供者可能依其設定與政策處理您的輸入內容、IP 位址、請求紀錄或其他網路中繼資料。
您應自行評估並負責所選伺服器、模型與 API 端點的隱私、安全性及資料保留政策。請勿向不受信任的伺服器傳送敏感或機密資訊。
上述連線模式說明，不代表 H3 裝置端影片、本地音樂生成或轉譜功能會將輸入傳送至自訂端點；這些功能仍在本地執行處理。
本 App 提供純文字形式的「回報」或「意見反應」功能。只有在您主動按下提交時，回報文字及隨附的裝置與系統資訊才會傳送至我們使用的開發者資料庫。本 App 不會自動上傳生成紀錄、轉譜紀錄或診斷資料。
App 內回報僅包含：
回報內容：您選擇提交的文字，例如意見、問題說明，或您刻意納入的 AI 文字與提示詞。此功能不接受圖片、影片、音訊或縮圖附件。
裝置與系統資訊：App 版本、裝置型號、Android 系統版本及系統語言設定，僅在您提交回報時一併傳送。
這些資料僅用於診斷問題、修復錯誤、提升 AI、圖片生成、影片生成、音樂生成與轉譜功能的穩定性、改善語言支援、回應您的回報，以及處理回報的濫用或安全問題。我們不會出售這些資料，也不會將其用於廣告、行銷、跨服務追蹤或建立廣告使用者檔案。
為營運回報資料庫、維護服務安全或遵守法律義務所必要時，相關資料可能由代表我們提供託管或基礎設施服務的供應商處理，或依法向有權機關揭露。此類處理僅限於必要目的。
在開放相關 H3 模型下載前，本 App 會詢問您是否位於限制地區，並依答案決定是否開放下載。答案僅在您的手機本地處理，不會因該聲明流程而傳送至我們的資料庫、Hugging Face、MiniMax 或其他第三方，亦不屬於 App 內回報隨附的診斷資訊。若您自行在回報文字中填寫所在地資訊，該提交文字依第 1.4 節處理。
本地聲明答案與模型來源在您連線下載時可能收到的 IP 位址及網路資訊不同；後者依第 4 節說明。
本地提示詞、來源圖片、影片與音訊、生成圖片、影片、音軌、音樂與轉譜結果、模型檔案、快取及其他裝置端資料，會依本 App 的功能、您的操作及 Android 系統設定保留在您的裝置上。您可透過本 App 提供的功能、Android 設定或檔案管理工具進行管理或刪除。清除 App 資料或移除本 App，可能使 App 私有儲存區內的資料被刪除。
已另行儲存、分享或匯出至本 App 外部的圖片、影片、音訊、轉譜結果或其他檔案，須由您在相應的儲存位置自行刪除。刪除 App 內副本，不一定會刪除已匯出的副本。
透過 App 內回報功能收到的文字回報及隨附裝置與系統資訊，將安全地儲存，並自收到之日起六（6）個月內自動且永久刪除。此保留期限不適用於僅保存在您裝置上的媒體或其他資料。
若適用法律要求我們保留較長期間，或為處理進行中的安全事件、法律請求或爭議所合理必要，我們可能在必要範圍內延長保留期間；相關目的結束後即予刪除。
依您所在地適用的資料保護法律，您可能有權要求查閱、更正、刪除、限制處理或反對處理您所提交的資料，並可在適用時撤回同意或向主管機關提出申訴。
如欲行使權利，請寄送電子郵件至 zawshishawn@gmail.com，並盡可能提供回報的大致日期與時間、裝置型號、App 版本、回報識別資訊或其他有助於定位紀錄的資料。我們只會要求為確認請求及查找相關紀錄所合理必要的資訊，並會在適用法律要求的範圍內處理您的請求。
由於本 App 不一定要求帳戶或固定身分識別資料，若無法合理辨識與您的請求相對應的紀錄，我們可能無法完成特定的查閱、更正或刪除要求。
本 App 未整合用於監控使用者行為或建立廣告使用者檔案的第三方廣告 SDK、跨服務追蹤軟體或分析工具，例如 Google Analytics 或 Firebase Analytics。
本 App 的功能（包括本地有聲影片生成）免費使用，不透過付款、訂閱或購買生成點數解鎖。若本 App 提供自願贊助，且您選擇透過 Google Play 贊助，交易由 Google 處理；Google 可能依其自身條款與隱私政策處理付款、Google 帳戶、裝置、購買及交易資訊。
就自願贊助交易，本 App 可能接收確認或處理交易所需的有限產品與購買狀態資訊，而非用於啟用付費功能或訂閱權益。我們不會直接取得您的完整付款卡片資料。贊助不影響功能存取。
本 App 可能使用 Hugging Face API 取得公開模型資訊，例如模型名稱、描述、中繼資料及下載連結。下載文字、圖片、影片或音樂生成模型、轉譜模型或相關元件時，您的裝置可能直接連線至 Hugging Face 或相關模型檔案來源。H3 相關功能會透過本 App，從 Hugging Face 下載由開發者提供的量化模型及相關元件。
Hugging Face 及相關模型檔案來源可能依其自身隱私政策，處理您的 IP 位址、請求紀錄、裝置或網路中繼資訊。因此，雖然生成與轉譜在本地進行，模型下載仍涉及第三方網路資料處理。下載完成的模型檔案會儲存在您的裝置上，直到您透過本 App、Android 設定或檔案管理工具將其刪除。
瀏覽模型清單或下載模型，不會使本 App 將本地提示詞、參考圖片、來源影片或音訊、生成圖片、影片、音軌、音樂、轉譜結果，以及縮圖上傳至 Hugging Face 或 MiniMax，也不會因此傳送本地地區聲明答案。若您另行使用第三方託管推論服務，該獨立處理依該服務的隱私政策辦理；H3 裝置端功能、本地音樂生成與轉譜均非此類服務。
https://huggingface.co/privacy
我們對透過回報功能收到的資料採取合理的技術與管理措施，以降低未經授權存取、遺失、濫用、變更或揭露的風險。然而，任何電子傳輸或儲存方式均無法保證絕對安全。
裝置端資料的安全性亦取決於您的裝置、作業系統、螢幕鎖定、備份設定及您選擇的自訂伺服器。對於由您控制或選擇的第三方伺服器，我們無法保證其安全性。
本 App 並非以 13 歲以下兒童為主要對象；若特定管轄區要求更高的最低年齡，例如 16 歲，則以該地法律規定為準。我們不會故意收集兒童的個人資料。
若您是家長或監護人，並認為兒童在未取得適當同意的情況下透過回報功能向我們提交資料，請立即透過 zawshishawn@gmail.com 與我們聯繫。我們將在能合理辨識相關紀錄的範圍內，儘速採取適當的刪除措施。
我們可能因功能、技術、法律或營運方式變更而不定期更新本隱私政策。更新後的版本將公布於本頁面，並修改「生效日期」。若變更對您的權利或資料處理方式具有重大影響，我們可能在適當情況下於本 App 內提供額外通知。
建議您定期查看本隱私政策，以瞭解最新內容。
若您對本隱私政策有任何疑問、建議，或希望行使資料權利，請透過以下方式聯絡我們：
開發者：趙士翔（Max Chao）
Email：zawshishawn@gmail.com