MIT 许可的桌面 AI 智能体工作台，可运行 DeepSeek-TUI 等 9 种本地引擎并统一配置 DeepSeek 模型路由；README 本身未直接提及 dsh/DeepSeek Harness。
DSH 适配
生态相关
作者声明
安全审计
未审计
最后核验
2026-08-21
许可证
MIT
01它能帮你完成什么？
Run DeepSeek-TUI and other local coding agents through a visual desktop chat instead of the terminal
Graphical chat with tool panels, slash commands, file diffs and permission prompts, plus a live workspace showing file writes, code changes and artifacts
Developers who use terminal agents such as DeepSeek-TUI, Claude Code or Codex and want a beginner-friendly GUI workspace on macOS or Windows
Configure DeepSeek and other model providers once and route them into the selected agent engine
A unified provider/model setup (official OpenAI, Anthropic, Gemini plus OpenAI-compatible endpoints for DeepSeek, Qwen, Moonshot, Ollama, OpenRouter, GitHub Copilot, local gateways) mapped into engines that follow WeSight settings
Users who switch between multiple LLM providers and want one place to manage model routing for their local agents
Monitor agent runs and extend workflows with skills and scheduled jobs
Runtime dashboard measuring engine, model, tokens, TTFT, TPS, tool latency, steps, status and duration; SkillHub skill installation and recurring agent tasks
Teams and hobbyists who run agent tasks daily and want metrics, skill management and scheduled automation in one desktop app
02如何接入 DeepSeek Harness？
前置条件
- macOS (Apple Silicon or Intel) or Windows x64 for the packaged desktop app
- For development from source: Node.js `>=24 <25` and npm
安装步骤
- 01
macOS: download the matching DMG from the latest GitHub Release (signed and notarized builds for Apple Silicon and Intel), open the DMG and drag `WeSight.app` into the `Applications` folder
- 02
Windows: download `WeSight.Setup.<version>.exe` from the latest release and check it against `SHASUMS256.txt` before running (the installer is unsigned and SmartScreen may warn)
- 03
From source (development): `git clone https://github.com/freestylefly/wesight.git`, `cd wesight`, `npm install`, then `npm run electron:dev`
$ git clone https://github.com/freestylefly/wesight.git
验证接入成功
作者未说明
03DSH 适配与能力边界
Not a dsh plugin: WeSight is a standalone desktop workspace that runs DeepSeek-TUI (DeepSeek's TUI agent runtime) as one of nine local engines and configures DeepSeek as a model provider; the README never mentions DeepSeek Harness/dsh itself
Agent engine manager
Local agent CLIs: Claude Code, Codex, Kimi Code, OpenClaw, Hermes Agent, OpenCode, Qwen Code, DeepSeek-TUI, or the built-in runtime→One-click install or detection of existing CLIs (macOS), reuse of existing CLI accounts/config, and visual chat execution
Can install local CLIs on the machine (macOS one-click setup)Tool execution runs commands and touches files, surfaced with permission promptsUnified model routing
Official OpenAI / Anthropic Claude / Google Gemini providers plus OpenAI-compatible endpoints (DeepSeek, Qwen, Moonshot, Ollama, OpenRouter, GitHub Copilot, local gateways, private endpoints)→Provider and model settings mapped into the selected engine, with switching between WeSight-managed settings and existing local CLI configuration
Model calls go over the network to the configured provider endpointsSettings and auth tokens are persisted in local SQLite under the app data directorySkillHub marketplace & scheduled tasks
SkillHub marketplace skills and WeSight built-in skills (research, documents, automation, creative, communication, agent building)→Install / enable / disable / update / remove skills from the desktop UI; recurring agent jobs for research, reports, monitoring and reminders
Skills are installed locally on diskScheduled tasks run agent jobs automatically on a recurring basisAI runtime telemetry dashboard
Agent runs across all configured engines→Per-task metrics: engine, model, source, status, tokens, completion time, TTFT, output-phase TPS, estimated model TPS, tool latency, agent steps
Runtime telemetry is persisted locally in SQLite
04适合谁？何时不该用？
适合
- Developers who use terminal agents such as DeepSeek-TUI, Claude Code or Codex and want a beginner-friendly GUI workspace on macOS or Windows
- Users who switch between multiple LLM providers and want one place to manage model routing for their local agents
- Teams and hobbyists who run agent tasks daily and want metrics, skill management and scheduled automation in one desktop app
不适合
- The Windows x64 installer currently has no code-signing certificate, so SmartScreen may show an unrecognized app warning; the README advises checking the file against `SHASUMS256.txt` before continuing.
05兼容性、维护与安全提示
- The Windows x64 installer currently has no code-signing certificate, so SmartScreen may show an unrecognized app warning; the README advises checking the file against `SHASUMS256.txt` before continuing.
- One-click installation of local agent CLIs is macOS-only; on other platforms you must bring an existing CLI setup, and packaged builds are published only for macOS (signed/notarized) and Windows x64 — no prebuilt Linux release.
- Running WeSight from source requires Node.js `>=24 <25` and npm — a narrow, pinned Node version range.
MIT · actively maintained (latest release v1.0.4, 2026-08-03)
06常见问题
如何在 WeSight 中使用 DeepSeek-TUI？
DeepSeek-TUI 是 WeSight 支持的 9 种引擎之一。macOS 上可以一键安装或自动检测已有安装；其他情况下在 Agent Engines 设置中选择它，并复用你现有的本地 CLI 账号和配置即可。
WeSight 原生运行 DeepSeek Harness（dsh）吗？
README 并未提及 dsh 本身。它与 DeepSeek 的关联是：把 DeepSeek-TUI 作为本地引擎运行（HTTP/SSE 运行时与工具流），以及在 WeSight 内把 DeepSeek 配置为 OpenAI 兼容的模型提供方。
支持哪些平台？
提供签名并公证的 macOS Apple Silicon 和 Intel DMG，以及 Windows x64 安装包（目前未签名，SmartScreen 可能警告，建议先核对 SHASUMS256.txt）。从源码开发需要 Node.js >=24 <25 和 npm。
我的数据存放在哪里？
本地数据（设置、会话、运行时调用、认证令牌）保存在应用数据目录下的 SQLite 中。模型调用会通过网络发往你配置的提供方，例如 DeepSeek、OpenAI、Anthropic、Gemini、Ollama 或 OpenRouter。
WeSight 能帮我安装智能体 CLI 吗？
macOS 上可以——WeSight 能一键安装受支持的本地 CLI，或检测机器上已有的安装；其他平台则需要你自带可用的终端环境，WeSight 会复用其账号和配置。
07相关的 DSH 工作流
open-design
作者 nexu-io
DeepSeek Harness 的开源 AI 设计插件，本地优先桌面应用，让编码智能体变身设计引擎，生成原型、落地页、幻灯片、图片与视频，支持多种格式导出。
picgo
作者 molunerfinn
PicGo：面向高效创作者的图片上传神器，支持 Obsidian、Typora、VS Code 等编辑器，以及 S3、GitHub、Cloudflare R2、Imgur、阿里云 OSS 等 60+ 图床，粘贴即上传。
dsh-desktop
作者 anywhere-labs
为 DeepSeek Harness (DSH) 插件生态打造的现代化桌面端解决方案。万物皆「插件」，桌面本身也是「插件」。
openbiliclaw
作者 whiteguo233
本地私有、开源的自进化跨平台 AI 内容发现 Agent：先理解你，再主动从 B站、小红书、抖音、YouTube、X、知乎、Reddit、微博等平台与开放 Web 寻找内容。（支持 deepseek harness 插件） | Local-first open-source cross-platform AI content discovery agent: understands you, then proactively finds content across Bilibili, Xiaohongshu, Douyin, YouTube, X, Zhihu, Reddit, Weibo and the open web.（support deepseek harness plugin）
08数据与来源
It helps you install or reuse Claude Code, Codex, Kimi Code, OpenClaw, Hermes Agent, OpenCode, Qwen Code, DeepSeek-TUI,…
Use existing Claude Code, Codex, Kimi Code, OpenClaw, Hermes Agent, OpenCode, Qwen Code, or DeepSeek-TUI accounts and co…
页面基于项目公开文档、仓库元数据和 DSH Plugins 的结构化解析生成；最后核验于 2026-08-21。发现错误？提交更正。
最佳 DeepSeek Harness 插件
从全目录挑出的 12 个值得优先安装的插件，覆盖各个分类。