Hermes Agent CN Desktop

Community desktop client for Hermes Agent, Windows-first, built with Tauri, with a bundled core and one-click install.

Stars
★ 1.7k
Last updated
28d ago
License
NOASSERTION
Primary language
TypeScript

At a glance

How it runs
Desktop app
Works with
Platform-specific
Cost
Free software; you pay for model usage
Setup effort
Medium · a few setup steps
You'll need
Node.js 20+pnpm 9+Rust stableHermes-CN-CoreShell / CLINetwork accessLocal filesystem
Typical use
Windows users who want a native desktop experience for Hermes Agent instead of the web Dashboard.
Not a fit if
  • Teams needing a stable release: current v0.8.0-rc4 is pre-release with changing APIs and UI
  • Commercial use or distribution: PolyForm NC 1.0.0 requires a separate paid licence
  • Linux-first users: official packages target Windows x64 and macOS only

What does this agent do, and when should you use it?

Hermes Agent CN Desktop is a desktop client from the Hermes Agent Chinese community, built with Tauri v2, Rust, React, and TypeScript, natively supporting Windows and macOS. It includes a community-modified Hermes Agent core (Hermes-CN-Core) and offers features like chat, streaming output, attachments, MCP tools, Skills, Memory, Profiles, scheduled tasks, LaTeX/Mermaid rendering, and a runtime health panel. The project focuses on desktop experience: native windows, local process management, file dialogs, managed runtime installation, runtime diagnostics, and a production-grade REST proxy layer with optional WS relay. Current version is v0.8.0-rc4, still under rapid development.

The repository acts as a desktop shell, using Tauri to provide native windows and manage the local Hermes Agent core's install, update, signature verification, health check, and rollback. The frontend is built with React and TypeScript, offering workbench, archive, chat, and task console views. It proxies REST and uploads via Rust commands, uses the official /api/ws for Gateway, and uses a Rust WS relay when needed to bypass WebView restrictions. It supports a wide range of model providers, including cloud services and local deployments (Ollama, vLLM, LM Studio, llama.cpp), and integrates with Feishu (Lark).

  1. Windows users who want a native desktop experience for Hermes Agent instead of the web Dashboard.
  2. macOS users who prefer a lightweight client that doesn't bundle Chromium, saving disk and memory.
  3. Developers who need to run a local Hermes Agent core and prefer a GUI for lifecycle management and diagnostics.
  4. Teams using Feishu and needing easy integration configuration in a desktop app.
  5. Power users who want to enable YOLO mode to auto-approve risky commands, with a toggle in settings.

How do you install or deploy this agent?

Download the installer for your platform from the official site (https://desktop.hermesagent.org.cn) or GitHub Releases (e.g., Windows x64 setup EXE, macOS Apple Silicon DMG). The installer includes a pre-bundled runtime for first-run initialization.

How do you use this agent?

After installation, configure an API key or local model endpoint. For development, follow the README: run pnpm install, start Hermes Dashboard with hermes dashboard --host 127.0.0.1 --port 9120 --no-open, then pnpm web:dev and cargo run to launch the desktop app in dev mode.

What are this agent's strengths and limitations?

Pros
  • Windows-first with macOS support; lightweight thanks to Tauri using the system WebView.
  • Bundles an independent Hermes Agent core with update, signature check, health check, and rollback.
  • Supports a broad range of model providers, including local solutions like Ollama and vLLM, and integrates Feishu.
  • Production-grade transport bridge: Rust command proxies REST and uploads, with centralized auth handling.
  • YOLO mode toggle for auto-approving dangerous commands, increasing automation efficiency.
Limitations
  • License is PolyForm Noncommercial 1.0.0; commercial use requires a separate license from Qingdao Wanderminds Network Technology Co., Ltd.
  • Early-stage project (v0.8.0-rc4); APIs, packaging, and UI are subject to change.
  • Depends on Hermes-CN-Core; requires separate installation or bundled runtime, adding complexity.
  • Platform support is primarily Windows x64 and macOS; Linux builds exist but are not officially highlighted.

How does this agent compare with similar options?

Compared to the stock Hermes Agent Dashboard (web), this project focuses on desktop-specific experience (native windows, process management, diagnostics) and includes a community-modified core (Hermes-CN-Core).

Key facts side by side with the most closely related agents.

Agent Source review Form / cost Stars Updated Language Full support on
Hermes Agent CN Desktop This agent 41 · Major gaps Desktop appFree + model costs ★ 1.7k 28d ago TypeScript —
Waku Coding Agent Desktop 69 · Some gaps Desktop appFree + model costs ★ 1.6k 16d ago Rust Codex · Claude Code
Xiaoma AI Hermes Chinese Enhancement Pack 36 · Major gaps — ★ 18 1mo ago Shell —
Polter Terminal Agent Supervisor 80 · Good Desktop appFree + model costs ★ 19 today Zig Claude Code

How does FollowAgents rate this agent?

FollowAgents source review · FARS-2.1
Major gaps
41/ 100 5-point scale 2.1 / 5
This review covers the source as of Aug 11, 2026. The project has shipped 1 stable release(s) since, which have not been re-reviewed. The score and verdict describe the reviewed version only. Recent releases ↓
Trust 8/29
Reliability 6/14
Adaptability 9/18
Convention 8/18
Effectiveness 7/13
Verifiability 3/8
Why each dimension lost points
Trust8 / 29 · 1.4/5

Evidence shows: project uses Tauri and Rust, has awareness of permission management (e.g., only necessary dependencies), but no explicit least-privilege design documentation. User confirmation mechanism not implemented. Data flow transparency partially described (e.g., transport bridge), but lacks detailed data flow diagrams. Sensitive data handling mentioned (e.g., credential leakage), but no specific measures shown. Dependency security has CI tests, but no vulnerability scanning. External effects mentioned (e.g., updates, downloads), but not detailed. Rollback mentioned (e.g., core rollback), but no specific implementation. Source attribution has author and community info, but publisher unverified. Deductions: lack of specific implementation evidence.

Reliability6 / 14 · 2.1/5

Evidence shows: project has CI tests (rust-test, web-test), but no test results provided. Dependency availability mentioned (e.g., bundled core), but no dependency list. Failure messages mentioned (e.g., error handling), but no specific examples. Deductions: lack of specific test results and error handling details.

Adaptability9 / 18 · 2.5/5

Evidence shows: project clearly targets Windows and macOS users, scenarios clear. Capability boundaries described (e.g., desktop shell vs core separation), but not detailed. Trigger precision mentioned (e.g., commands), but no specific trigger conditions. Environment fit mentioned (e.g., cross-platform), but no detailed configuration. Deductions: insufficient description of capability boundaries and trigger precision.

Convention8 / 18 · 2.2/5

Evidence shows: information architecture clear (repo layout), install notes detailed (quick start), naming stable (version numbers), examples and FAQ mentioned (e.g., screenshots), known limitations mentioned (e.g., unsigned), license clear (PolyForm NC), versioning changelog mentioned (e.g., release process), maintenance responsibility mentioned (e.g., community). Deductions: examples and FAQ content insufficient, versioning changelog lacks specific content.

Effectiveness7 / 13 · 2.7/5

Evidence shows: output usability mentioned (e.g., UI screenshots), marginal value mentioned (e.g., desktop experience), cost-benefit mentioned (e.g., lightweight). Deductions: lack of specific data support.

Verifiability3 / 8 · 1.9/5

Evidence shows: claim traceability mentioned (e.g., sources), cross-source corroboration mentioned (e.g., community), fact-inference separation mentioned (e.g., descriptions). Deductions: lack of specific verification evidence.

Risks and how to mitigate them
  • Not found in source: confirmation before actingTurn on (or add) a confirmation step before it acts, and try it in a sandbox or test environment before real data.
  • Publisher unverified, identity unknown.
  • Windows installer not Authenticode-signed, may trigger SmartScreen warnings.
  • License is PolyForm NC, commercial use requires separate authorization.
  • Project in alpha stage, APIs and features may change.
Evidence confidence: Low Reviewed Aug 11, 2026 Reviewed revision 8e92a903855f New commits since this review; the score may not cover them
See the full review method →

Recent releases

Latest entries from the project's GitHub Releases. Notes are quoted from upstream as written, not rewritten or tested by FollowAgents.

  1. v0.9.1-prototype.nightly.20260921.1 Hermes Desktop 0.9.1-prototype.nightly.20260921.1 Pre-release
    Release notes excerpt
    自动 Nightly,仅供 prototype/primelab-win 热更新验证。
    View on GitHub ↗
  2. v0.9.0 Hermes Agent 中文社区桌面版 v0.9.0
    Release notes excerpt
    Hermes Agent 中文社区桌面版 v0.9.0 本版升级至基于上游 v0.21 的中文社区内核,并带来统一的软件更新体验。 主要改进 - 软件更新更清楚:发现新版后通过弹窗和右下角状态栏提醒;查看更新说明后自行下载,下载期间可以继续使用,准备就绪后再决定何时生效。 - 保护正在进行的任务:更新涉及重启时,先等待受影响任务结束;“稍后提醒”会保留更新入口,24 小时内不重复提示同一版本。 - 保留高级功能:更新渠道、nightly、自定义更新源、内测配置、组件更新与回退集中到默认收起的“高级更新选项”。 - 修复关键问题:修复 Windows 网关启动、聊天历史备份、切换档案后离线或无法续聊、相同回复跨轮合并、澄清选项,以及冻结内核中的工具发现与终端资源问题。首次模型配置入口现在会正确选中推荐的服务商。 安装与升级 - macOS:需要 macOS 14 或更高版本,请按 Apple Silicon 或 Intel 选择安装包。正式安装包经过 Developer ID 签名和 Apple 公证。[历史版本下载](https://github.com/Eynzof/Hermes-CN-Desktop/releases/tag/v0.7.0)继续保留。 …
    View on GitHub ↗
  3. v0.7.0 Hermes Agent 中文社区桌面版 v0.7.0

FAQ

What are the terms for commercial use?
Non-commercial use is under PolyForm Noncommercial License 1.0.0. For commercial use, distribution, or integration, you must obtain a separate license from Qingdao Wanderminds Network Technology Co., Ltd. Contact: [email protected].
Does it support Linux?
The repo can build Linux .deb and .AppImage packages, but the project officially targets Windows and macOS; full Linux support is not explicitly documented.
How do I update the core?
The desktop app supports managed runtime updates and rollbacks. Specific steps are not detailed in the README; refer to runtime diagnostics and documentation.
Which model providers are supported?
It supports major cloud providers and local deployments like Ollama, vLLM, LM Studio, and llama.cpp, plus Feishu integration.
View on GitHub ↗ Install ↓

Related agents