BoxPlayer v5.0.23
中文
AI 工作台与文档理解
- AI 工作台升级为更可靠的任务执行链路:文件检索、媒体获取与工作区操作会在主进程中统一编排,并保留可追溯的执行记录。
- 文档 AI 支持在文件预览中发起连续追问;回答可关联本次使用的来源,便于回看与核对。
- AI 搜索与会话状态同步优化,减少任务完成后界面状态不同步的问题。
媒体与刮削
- 优化视频媒体刮削:临时网络错误、限流和服务端错误会按退避策略重试,不再被误判为“未匹配”。
- 加强文件名清洗、识别诊断与未匹配分析;仅上报改善识别率所需的非敏感诊断信息。
- 统一外链、磁力、分享和字幕获取的执行流程,提升任务恢复和错误提示的一致性。
网盘与上传
- 修复拖拽上传到非阿里云盘子目录时,因目录 ID 格式校验而被拒绝的问题。
- 修复 Dropbox 子目录上传丢失目标路径上下文的问题;文件和文件夹任务均会保存目标目录信息。
- 强化百度网盘上传目标路径解析,避免拖拽后因页面状态变化而上传到错误位置。
- 改善多个网盘的列表/文件模型兼容性与上传任务持久化。
稳定性与分析
- 修复应用启动时分析服务初始化时机,确保错误与使用事件能够正常发送。
- 增加媒体扫描、AI 刮削和上传目标的回归测试。
English
AI Workspace and document insights
- The AI Workspace now uses a more reliable task execution pipeline. File search, media acquisition, and workspace actions are orchestrated in the main process with an auditable execution history.
- Document AI supports follow-up questions from file preview, with source-aware responses for easier review.
- Improved AI Search and session state synchronization after task completion.
Media and metadata matching
- Video matching now retries transient network, rate-limit, and server failures with backoff instead of incorrectly marking them as unmatched.
- Improved filename normalization, diagnostics, and unmatched-media analysis using only the non-sensitive data required to improve recognition.
- Unified the execution flow for external URLs, magnets, shares, and subtitles for more consistent recovery and error reporting.
Cloud drives and uploads
- Fixed drag-and-drop uploads being rejected in non-Aliyun subfolders because of an Aliyun-specific folder-ID validation rule.
- Fixed Dropbox subfolder uploads losing their destination-path context; both file and folder tasks now retain the target folder information.
- Hardened Baidu upload destination resolution so drag uploads cannot be redirected by later UI state changes.
- Improved cross-provider file-list compatibility and upload-task persistence.
Reliability and analytics
- Fixed application-start analytics initialization so usage and error events are sent reliably.
- Added regression coverage for media scanning, AI matching, and upload target handling.