Wenanbao
01 Scope and operator
- Operator
- To be completed before public release
- Address
- To be completed before public release
- Privacy email
- To be completed before public release
- Support email
- To be completed before public release
This statement covers information handling in the Wenanbao mobile app and this static website. The product is in development testing without a public store release. Changing the interface language does not change the service region, backend or speech provider.
02 Local recording and basic use
Audio, imported audio copies, titles, durations, markers, transcripts, edits, document history and interface preferences are stored in app-managed device storage. There is no built-in account, cloud backup or cross-device sync. System backups, device migration and sharing follow your device settings or the recipient you select.
You can record, play back and manage local content without cloud transcription. Local recording does not actively read contacts, precise location, SMS, call logs or advertising identifiers. Call recording and voiceprint identity authentication are not provided.
03 When cloud transcription happens
After you select a recording and confirm transcription in the app, the selected audio, recognition language, task identifier and necessary processing information are sent to the configured backend. It temporarily stores or converts the audio, makes it available to Volcengine speech recognition, and receives text, timing and speaker-segment labels. Those labels do not identify participants by their real names.
Local recording does not automatically upload audio to a model service. Current AI text organization is a demonstration, not a live text-model upload. We do not sell your content or use it for advertising profiles, and have not implemented training our own models on user content.
04 Permissions and device state
Microphone access is requested for recording. Without it, audio capture is unavailable; existing documents and available imports remain accessible. Android notifications display recording state and controls. Hiding them does not mean recording has stopped. A foreground service and wake lock support capture while recording; they are not used to start recording on boot.
The Android background checklist reads microphone permission and mute state, app and recording-channel notification settings, standard battery optimization, Battery Saver, background restrictions and manufacturer name for relevant guidance. It does not read call numbers or content, and does not change power settings for you. File import and sharing use system pickers for content you select.
05 Service providers and components
The current test transcription integration uses Volcengine (Beijing Volcano Engine Technology Co., Ltd.) to turn audio you confirm into text. It involves the selected audio, recognition language, service task identifiers and recognition results. Access is through backend APIs; the mobile app does not integrate Volcengine advertising or growth-analytics SDKs.
Provider policy: Volcengine privacy policy.
Flutter and components including record, path_provider, shared_preferences, file_picker, audioplayers and share_plus support the interface, recording, storage, preferences, file selection, playback and system sharing. The http component handles explicitly initiated backend requests. No advertising or third-party behavioral analytics service is integrated. Provider retention and training-use terms must be checked against the actual service agreement before public release; this statement does not establish the provider’s retention behavior.
- Provider retention
- To be completed before public release
- Provider training use
- To be completed before public release
06 Retention and deletion
Device records remain until you delete them or clear app data. Deletion attempts to remove the audio, recovery journal, repaired-tail copy and validation cache. Uninstalling normally removes local app data; system-backup copies follow system settings. Exported copies are managed by you and the recipients.
Under the current backend’s normal operation, source files are removed after conversion; temporary recognition audio is cleaned up when a task succeeds, fails or expires, with a maximum processing deadline of four hours. Results and task metadata are cleaned up about 24 hours after the terminal state. Downtime or failed deletion can delay cleanup, with retries after recovery. This does not describe the speech provider’s retention period.
Deleting a local record does not automatically cancel a submitted cloud task or instantly remove provider-side copies. To request handling of retained cloud content, use the privacy contact on this page and provide task information. Do not resend unrelated full recordings with your request.
07 Hosting region and network information
- Backend region
- To be completed before public release
- Hosting security-log retention
- To be completed before public release
The current test setup uses Volcengine’s China-region transcription configuration. Overseas services are not live. An English interface or website does not imply overseas processing. Services for new regions must disclose actual recipients, hosting regions and processing rules and complete applicable requirements before enabling uploads.
Website access, backend requests and temporary audio downloads involve connection information such as IP address, connection time and request path. Generic backend access logging is off by default. Actual hosting, proxy or CDN security logging must be disclosed for production and limited to service operation and security.
08 Website preferences
This static website has no audio or document upload, advertising or analytics cookies, or external font requests. Selecting an appearance attempts to store only a light, dark or system theme preference in your browser. Browsing still works if storage is blocked. Clear site data to remove it. External links open when selected, and the destination handles access under its own rules.
09 Your choices and rights
You can decline cloud transcription, revoke microphone and notification permissions in system settings, edit or delete local records, and export available text and audio. Revoking permission does not automatically delete history or reverse completed processing. Contact us to request access, correction, copies, deletion or to object to processing. Verification will be limited to what is needed for the request, with a response within applicable legal time limits.
10 Sensitive content and minors
Recordings may contain other people’s personal or sensitive information. Before recording, importing or uploading, ensure you are entitled to process it and have required authorization. The test version has no child-specific service or children’s-data workflow and does not offer cloud recording processing to children under 14. Contact us if children’s information has been submitted in error.
11 Security, updates and contact
The app uses private storage, atomic recovery journals and complete-frame audio checks. Temporary backend audio downloads use time-limited signed URLs. Production network services must use HTTPS. Storage and recovery measures cannot guarantee lossless recovery after device damage, power loss or storage failure; export important content.
Changes to purposes, data scope, recipients or material rules require an updated statement and, where needed, renewed notice or consent. The version, effective status and date are stated at the top of this page. Contact details appear in section 1. Public cloud services should not open before contact channels are completed.