Skip to content

ClawdHome 1.12.0

Latest

Choose a tag to compare

@deepjerry-ai deepjerry-ai released this 06 Jun 18:01

中文

重点更新

  • 新增 Hermes Web UI 视图与管理器,方便查看和管理 Hermes 运行状态。
  • AI 实验室新增完整语音转写体验,支持转写、编辑、历史记录、导出与保存,并显示进度阶段和实时转写速度。
  • 虾塘各实例(OpenClaw 与 Hermes)新增数字形象展示,内置多组团队与角色头像素材,让实例识别与管理更直观。
  • AI 对话已接入全局模型账户,并支持图片输入。
  • 模型管理新增本地模型下载器,全局模型池与提示词记忆同步升级。

体验优化

  • 语音转写支持原生 macOS 语音隔离与降噪,并修复部分音频场景下 ASR 重复幻觉的问题。
  • 仪表盘新增本机整体网络吞吐趋势,实时展示入站与出站速率,排查网络活动更直观。
  • 语音转写设置页新增内置 ASR 工具状态诊断,可查看可用性、版本、构建时间与失败原因。
  • 设置页权限指引更精简,辅助功能与 Chrome 自动化说明仅在需要时展开。
  • 仪表盘、用户列表、设置、备份、权限与文件管理器界面继续优化。
  • 受管浏览器启动不再需要主机辅助功能权限。

English

Highlights

  • New Hermes Web UI view and manager for easier Hermes status review and management.
  • AI Lab now includes a full speech transcription experience with transcription, editing, history, export/save, progress stages, and real-time speed display.
  • Shrimp Pool instances across OpenClaw and Hermes now include digital avatars, with new team and persona avatar assets for clearer instance recognition and management.
  • AI chat now uses global model accounts and supports image input.
  • Model management now includes a local model downloader, plus upgrades to the global model pool and prompt memory.

Improvements

  • Speech transcription now supports native macOS vocal isolation and noise suppression, and fixes repeated ASR hallucinations in some audio scenarios.
  • The dashboard now shows whole-machine network throughput trends with live inbound and outbound rates.
  • Speech transcription settings now show bundled ASR tool diagnostics, including availability, version, build time, and failure details.
  • Settings permission guidance is quieter, with Accessibility and Chrome Automation instructions expanded only when needed.
  • Dashboard, user list, settings, backup, permissions, and file manager UI polish continues.
  • Managed browser launch no longer requires host Accessibility permission.