Skip to content

LingShu v0.1.0-alpha.4

Pre-release
Pre-release

Choose a tag to compare

View all tags
@RoyZhao1991 RoyZhao1991 released this 17 Jul 22:46
v0.1.0-alpha.4
52ab1cc

LingShu v0.1.0-alpha.4

English | 简体中文

English

This alpha makes the first-run language choice consistent from the first visible message while preserving the agent, CLI, human-collaboration, and artifact-delivery behavior introduced in the earlier public alphas.

What changed

  • English first launch now starts with an English Nous greeting instead of the previous hard-coded Chinese message.
  • Clearing the main conversation creates a new greeting and execution-trace reset message in the currently selected language.
  • Switching the first-run language updates only the untouched welcome message. Existing conversation content is never rewritten.
  • The language preference and highest-priority model instruction now share one small support boundary instead of expanding the UI localization module.

Core capabilities

  • Turn a report brief into real editable presentation or document files, register and preview them locally, revise the actual output, and hand it to an independent checker.
  • Bring your own model through OpenAI, Anthropic Claude, DeepSeek, MiniMax M3, or a custom compatible endpoint.
  • Convert requests into structured goals, run mutable workflows through isolated workers and local tools, and verify results before completion.
  • Pause an exact worker or checker session for typed human collaboration such as login, QR scanning, file selection, confirmation, or physical action, then resume from that checkpoint.
  • Use native macOS Accessibility-based Computer Use without requiring Codex.
  • Use the bundled lingshu CLI as a one-request/one-response entrance for Feishu bots, webhooks, Shortcuts, and local automation while preserving the same serialized main session.

Verification and release integrity

  • Full local suite: 1,578 tests executed, 3 environment-dependent tests skipped, 0 failures.
  • GitHub CI passed on release source commit 52ab1cc2f184f50e8ac9780a5ecfa77383cd1c2f.
  • Build 7 was compiled from a clean source tree in isolated Apple silicon and Intel scratch directories.
  • The app and DMG were separately accepted by Apple notarization and carry stapled tickets.
  • The release script remounted the DMG and revalidated the installed app, bundled CLI, signature, notarization, and embedded source revision.
  • DMG SHA-256: daaf69d7eb1e5fc76d6fdf9a9c2216a3ba87aea7b90fef4c30fdff3b647b2518.

Install

With Homebrew:

brew install --cask RoyZhao1991/tap/lingshu

Or download the Universal DMG and .sha256 file from this release, then verify it:

shasum -a 256 -c LingShu-0.1.0-7-macOS-universal.dmg.sha256

Open the DMG, drag 灵枢.app to Applications, launch it, choose a language, and follow the brain-channel setup guide. Grant macOS permissions only when a capability you choose requires them.

Requirements and limitations

  • macOS 14 or later.
  • A token for a supported model provider, or a compatible custom endpoint.
  • LingShu is alpha software. Keep backups and review high-impact actions and generated deliverables.
  • Remote model and perception requests follow the configured provider's privacy and retention terms.
  • This DMG does not bundle the optional SenseVoice model runtime or experimental HAL virtual microphone driver; Apple Speech remains available as a local fallback.

简体中文

本次 Alpha 让首次语言选择从第一条可见消息开始保持一致，同时保留此前公开版本已经提供的 Agent、CLI、人机协作与真实产物交付能力。

本次改动

  • 首次选择英文后，欢迎语现在会以 Nous 和英文显示，不再出现写死的中文首条消息。
  • 清空主对话后，新欢迎语和执行轨迹重置提示会使用当前所选语言。
  • 首次切换语言只更新尚未发生真实对话的欢迎语，不会重写已有会话内容。
  • 语言偏好与模型最高优先级语言指令归入独立的小型支持模块，避免继续扩大界面国际化模块。

核心能力

  • 把一条汇报需求推进为真实可编辑的 PPT 或文档，在本地登记、预览和修改，并交给独立 checker 验收。
  • 支持 OpenAI、Anthropic Claude、DeepSeek、MiniMax M3 与自定义兼容端点，主脑可替换。
  • 将请求整理为结构化目标，通过可变工作流、隔离 worker 和本地工具推进，并在完成前验收。
  • worker 或 checker 可因登录、扫码、选文件、确认或实体操作暂停精确会话，用户完成后从原断点继续。
  • 通过 macOS 原生辅助功能接口进行 Computer Use，不依赖 Codex。
  • 随 App 交付 lingshu CLI，为飞书机器人、Webhook、快捷指令和本地自动化提供一问一答入口，并复用同一串行主会话。

验证与发布完整性

  • 本地完整回归：执行 1,578 项测试，3 项环境相关测试按设计跳过，0 失败。
  • 发布源码提交 52ab1cc2f184f50e8ac9780a5ecfa77383cd1c2f 的 GitHub CI 已通过。
  • Build 7 从干净工作区构建，Apple 芯片与 Intel 分别使用隔离的临时编译目录。
  • App 与 DMG 分别通过 Apple 公证并装订公证票据。
  • 发布脚本重新挂载 DMG，再次校验其中 App、内置 CLI、签名、公证状态和源码提交号。
  • DMG SHA-256：daaf69d7eb1e5fc76d6fdf9a9c2216a3ba87aea7b90fef4c30fdff3b647b2518。

安装

使用 Homebrew：

brew install --cask RoyZhao1991/tap/lingshu

也可以下载本 Release 的 Universal DMG 与 .sha256 文件并校验：

shasum -a 256 -c LingShu-0.1.0-7-macOS-universal.dmg.sha256

打开 DMG，把 灵枢.app 拖入“应用程序”，启动后选择语言并按引导配置主脑。仅在所选能力确实需要时授予对应 macOS 权限。

环境要求与限制

  • 需要 macOS 14 或更高版本。
  • 需要受支持模型服务商的 Token，或兼容的自定义端点。
  • 灵枢仍处 Alpha 阶段，请保留重要文件备份，并检查高影响操作与生成产物。
  • 发送到远程模型或感知服务的内容受对应服务商隐私和留存条款约束。
  • 本 DMG 不内置可选的 SenseVoice 模型运行时或实验性 HAL 虚拟麦克风驱动；Apple Speech 仍可作为本地回退方案。
Assets 5
Loading