Releases
v1.13.0
Compare
Sorry, something went wrong.
No results found
中文
重点更新
AI 实验室新增“内容本地脱敏”:可在本机识别敏感实体,生成可审阅的脱敏文本,支持映射还原与 Markdown 导出。
语音转写支持流式结果、历史同步、JSON/SRT 导出和可调的块级时间戳。
内置 Node.js 运行时升级至 26.5.0。
稳定性与兼容性
限制长音频转写的 MLX 内存占用,并补充缓存管理与内存遥测信息。
改进 ASR 架构检测、音频访问提示与工具错误报告,提升不同 Mac 环境下的诊断体验。
持续优化语音历史去重、Obsidian 同步和浏览器会话相关交互。
English
Highlights
AI Lab now includes Content Redaction: identify sensitive entities on-device, review redacted text, restore mappings, and export Markdown.
Speech transcription now supports streamed results, history sync, JSON/SRT export, and adjustable chunk-level timestamps.
The bundled Node.js runtime is updated to 26.5.0.
Stability and Compatibility
Bound MLX memory use for long audio transcription, with improved cache management and memory telemetry.
Improve ASR architecture detection, audio-access guidance, and tool error reporting across Mac environments.
Continue polishing speech history deduplication, Obsidian sync, and browser-session interactions.
You can’t perform that action at this time.