SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillshermes-remote-deploy

hermes-remote-deploy

Deployment & operations

View repository

Updated 10-15 12:33

About this skill

Hermes Remote Deploy

Quickly deploy Hermes Agent on a remote Linux/macOS server via SSH — a self-evolving AI Agent from Nous Research, with a built-in learning loop, multi-platform gateway, and skill system.

What is Hermes: It is not a coding copilot tied to an IDE, nor a chat shell for a single API. It is an autonomous Agent running on your server that remembers what it learns and gets stronger the more you use it. It supports multiple terminal backends such as Docker / SSH / Modal / Daytona / Vercel Sandbox, and connects through multiple platforms such as Telegram / Discord / Slack / WhatsApp / Signal.

Prerequisites

Before deployment, ensure:

  • Control side: Already through ssh-copy-id added the public key to the target machine
  • Target machine: Linux (Ubuntu/Debian/CentOS) or macOS, with SSH service enabled
  • Target machine: curl and git already installed (basic tools)
  • Keys: Hold an API Key for at least one model provider (such as DeepSeek / OpenAI / Anthropic / OpenRouter, etc.)
  • Network: The target machine can access GitHub (github.com) and PyPI (pypi.org）

Note: The installation script automatically installs all dependencies including Python 3.11, Node.js 22, uv, ripgrep, ffmpeg, etc.; no manual pre-installation is required.

Workflow

Deployment is divided into 5 steps, automatically completed by scripts/deploy.sh :

Step Operation Description
1 Connectivity check Verify SSH reachability and detect the target system type
2 Install Hermes curl one-click installation script, automatically handles all dependencies
3 Configure model hermes config set Configure Provider and API Key
4 Configure identity Optional: deploy SOUL.md and AGENTS.md
5 Verify hermes doctor diagnostics + hermes --version

Usage

Quick execution

bash scripts/deploy.sh <user@host> <provider> <api-key>

Example：

# DeepSeek 作为默认模型
bash scripts/deploy.sh daniel@my-vps-01 deepseek sk-2312xxxx

# OpenAI
bash scripts/deploy.sh daniel@my-vps-01 openai sk-proj-xxxx --model openai/gpt-4o

# Anthropic
bash scripts/deploy.sh daniel@192.168.1.100 anthropic sk-ant-xxxx

# OpenRouter (支持 200+ 模型)
bash scripts/deploy.sh daniel@my-vps-01 openrouter sk-or-v1-xxxx --model openrouter/anthropic/claude-sonnet-4

# 预览模式（不执行）
bash scripts/deploy.sh daniel@my-vps-01 deepseek sk-xxxx --dry-run

# 跳过 Gateway 配置（纯 CLI 模式）
bash scripts/deploy.sh daniel@my-vps-01 deepseek sk-xxxx --skip-gateway

Parameters

Parameter Required Default value Description
<user@host> ✅ — SSH target, such as root@my-vps or user@192.168.1.100
<provider> ✅ — Provider name:deepseek / openai / anthropic / openrouter / nous
<api-key> ✅ — API Key corresponding to the Provider
--model No Provider default model Model ID (such as deepseek/deepseek-v4-pro）
--skip-gateway No false Skip Gateway configuration and deploy only the CLI
--soul-path No — Local SOUL.md path, deployed to remote
--branch No main Hermes installation branch
--dry-run No false Only print commands, do not actually execute

Provider default model

Provider Default model API Key environment variable
deepseek deepseek/deepseek-v4-pro DEEPSEEK_API_KEY
openai openai/gpt-4o OPENAI_API_KEY
anthropic anthropic/claude-sonnet-4 ANTHROPIC_API_KEY
openrouter openrouter/anthropic/claude-sonnet-4 OPENROUTER_API_KEY
nous nous/hermes-3-70b NOUS_API_KEY

Manual step-by-step execution

If the automatic script is not applicable, you can manually operate according to the following steps:

Step 1: Connect to the target machine

ssh user@target-host

Verify:echo SSH_OK && uname -a

Step 2: Install Hermes

# 一键安装（国内可先设置代理）
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash

# 或指定分支
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash -s -- --branch main

After installation is complete, reload the Shell:

source ~/.bashrc  # 或 source ~/.zshrc
hermes --version

Mainland China network tip: If GitHub access is slow, you can set up a proxy first export https_proxy=http://proxy:port, or download on a machine with a proxy and transfer it via scp.

Step 3: Configure the model Provider

# 设置模型（API key 自动存入 ~/.hermes/.env）
hermes config set model deepseek/deepseek-v4-pro
hermes config set DEEPSEEK_API_KEY sk-xxxx

# 查看当前配置
hermes config

For all supported Providers and configuration methods, seeofficial documentation。

Step 4: Run interactive setup (optional but recommended)

hermes setup

The interactive wizard will guide you through:

  • Model provider selection
  • Tool enable/disable
  • Gateway platform connection (Telegram / Discord / Slack, etc.)
  • Memory system preferences

Step 5: Configure identity and context files (optional)

# 部署 SOUL.md（Agent 人格）
cat > ~/.hermes/SOUL.md << 'EOF'
# Your Agent Identity
...
EOF

# 部署 MEMORY.md（跨会话记忆）
mkdir -p ~/.hermes/memories
cat > ~/.hermes/memories/MEMORY.md << 'EOF'
# Memory
...
EOF

Step 6: Start and verify

# 诊断检查
hermes doctor

# 启动 CLI 聊天
hermes

# 启动 Gateway（多平台接入）
hermes gateway setup   # 首次配置
hermes gateway start   # 启动

From OpenClaw Migration

If the target machine already has OpenClaw, Hermes supports one-click migration:

# 交互式迁移（自动检测 ~/.openclaw）
hermes claw migrate

# 预览模式
hermes claw migrate --dry-run

# 仅迁移用户数据（不含密钥）
hermes claw migrate --preset user-data

Migration contents include: SOUL.md, MEMORY.md, USER.md, Skills, API Keys.

Post-Deployment Checklist

  • hermes --version Output the version number
  • hermes doctor No critical errors
  • hermes config Display the correct model configuration
  • Can start CLI conversation normally:hermes
  • (Optional) Gateway platform connection is normal
  • (Optional) SOUL.md / MEMORY.md contents are correct

Troubleshooting

SSH connection failed

# 检查连通性
ssh -v user@host

# 确认目标机器 SSH 服务已开
# Linux: systemctl status sshd
# macOS: 系统设置 → 通用 → 共享 → 远程登录

# 确认公钥已复制
ssh-copy-id user@host

Installation script download failed

# 国内网络可尝试
export https_proxy=http://your-proxy:port
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash

# 或手动克隆
git clone https://github.com/NousResearch/hermes-agent.git ~/.hermes/hermes-agent
cd ~/.hermes/hermes-agent
uv sync --extra all

hermes command not found

# 重载 shell 配置
source ~/.bashrc  # 或 source ~/.zshrc

# 检查 PATH
echo $PATH | grep hermes

# 手动添加（安装脚本通常自动处理）
export PATH="$HOME/.hermes/hermes-agent/.venv/bin:$PATH"

Model API call failed

# 检查 API Key 配置
hermes config | grep -i key

# 测试连接
hermes doctor

# 手动设置 API Key
hermes config set DEEPSEEK_API_KEY sk-xxxx
# 或直接编辑 .env
vim ~/.hermes/.env

Special handling for mainland China environment

# 设置 uv 镜像
export UV_INDEX_URL=https://pypi.tuna.tsinghua.edu.cn/simple

# 设置 npm 镜像
export NPM_CONFIG_REGISTRY=https://registry.npmmirror.com

# 或在安装脚本前设置
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | \
  env UV_INDEX_URL=https://pypi.tuna.tsinghua.edu.cn/simple bash

Directory Structure (Post-Deploy)

~/.hermes/
├── config.yaml        # 主配置文件
├── .env               # API Keys 和密钥
├── auth.json          # OAuth 凭证
├── SOUL.md            # Agent 人格定义
├── memories/          # 跨会话记忆
│   ├── MEMORY.md
│   └── USER.md
├── skills/            # Agent 自创技能
├── cron/              # 定时任务
├── sessions/          # Gateway 会话
└── logs/              # 日志

Comparison with OpenClaw deployment

Dimension Hermes OpenClaw
Language stack Python 3.11 + Node.js 22 Node.js
Package management uv (Python) + npm (Node) npm
Installation method curl pipe bash npm install -g
Configuration method hermes config set openclaw config set
Configuration directory ~/.hermes/ ~/.openclaw/
Gateway Built-in (Telegram/Discord/Slack/WhatsApp/Signal) Built-in
Terminal backend 7 types (local/docker/ssh/modal/daytona/vercel/singularity) local
Learning loop ✅ Built-in skills system + memory system Skills system
Migration support ✅ hermes claw migrate —