curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash

如果网络报错设置代理

#设置PROXY 当前终端有效
export ALL_PROXY=socks5://192.168.88.1:7890
export HTTP_PROXY=socks5://192.168.88.1:7890
export HTTPS_PROXY=socks5://192.168.88.1:7890

设置界面

How would you like to set up Hermes?
↑↓ navigate ENTER/SPACE select ESC cancel

→ (●) Quick setup — provider, model & messaging (recommended)
(○) Full setup — configure everything

选择Quick setup回车

Select provider:
↑↓ navigate ENTER/SPACE select ESC cancel

(○) Nous Portal (Nous Research subscription)
(○) OpenRouter (100+ models, pay-per-use)
(○) Anthropic (Claude models — API key or Claude Code)
(○) OpenAI Codex
(○) Xiaomi MiMo (MiMo-V2 models — pro, omni, flash)
(○) Qwen OAuth (reuses local Qwen CLI login)
(○) GitHub Copilot (uses GITHUB_TOKEN or gh auth token)
(○) GitHub Copilot ACP (spawns `copilot --acp --stdio`)
(○) Hugging Face Inference Providers (20+ open models)
(○) Google AI Studio (Gemini models — OpenAI-compatible endpoint)
(○) Google Gemini via OAuth + Code Assist (free tier supported; no API key needed)
(○) DeepSeek (DeepSeek-V3, R1, coder — direct API)
(○) xAI (Grok models — direct API)
(○) Z.AI / GLM (Zhipu AI direct API)
(○) Kimi Coding Plan (api.kimi.com) & Moonshot API
(○) Kimi / Moonshot China (Moonshot CN direct API)
(○) MiniMax (global direct API)
(○) MiniMax China (domestic direct API)
(○) Alibaba Cloud / DashScope Coding (Qwen + multi-provider)
(○) Ollama Cloud (cloud-hosted open models — ollama.com)
(○) Arcee AI (Trinity models — direct API)
(○) Kilo Code (Kilo Gateway API)
(○) OpenCode Zen (35+ curated models, pay-as-you-go)
(○) OpenCode Go (open models, $10/month subscription)
(○) Vercel AI Gateway (200+ models, pay-per-use)
(○) AWS Bedrock (Claude, Nova, Llama, DeepSeek — IAM or API key)
(○) Custom endpoint (enter URL manually)
(○) Cancel

选择第一个 Nous Portal (Nous Research subscription) 回车

#提示
To continue:
1. Open: https://portal.nousresearch.com/device/verify?user_code=MEQW-AEGQ
2. If prompted, enter code: MEQW-AEGQ
Waiting for approval (polling every 1s)...

可用邮箱注册也可以用github gmail登陆
Subscription选择Free $0/mo 免费套餐

订阅需要使用strip支付通道验证
可以用visa 银联 AE卡

Approve Hermes Agent Login
Enter the code shown in your CLI, then approve or deny the login request.

User Code

这里usercode 和上面2的控制台一致即可

Waiting for approval (polling every 1s)...

Login successful!
Auth state: /home/hello/.hermes/auth.json

Showing 1 curated models — use "Enter custom model name" for others.
Select default model:
In Out Cache /Mtok

anthropic/claude-opus-4.6 $5.00 $25.00 $0.50
anthropic/claude-sonnet-4.6 $3.00 $15.00 $0.30
anthropic/claude-sonnet-4.5 $3.00 $15.00 $0.30
anthropic/claude-haiku-4.5 $1.00 $5.00 $0.10
openai/gpt-5.4 $2.50 $15.00 $0.25
openai/gpt-5.4-mini $0.75 $4.50 $0.07
openai/gpt-5.3-codex $1.75 $14.00 $0.17
google/gemini-3-pro-preview
google/gemini-3-flash-preview $0.50 $3.00 $0.05
google/gemini-3.1-pro-preview $2.00 $12.00 $0.20
google/gemini-3.1-flash-lite-preview $0.25 $1.50 $0.02
qwen/qwen3.5-plus-02-15 $0.26 $1.56
qwen/qwen3.5-35b-a3b $0.16 $1.30
stepfun/step-3.5-flash $0.10 $0.30
minimax/minimax-m2.7 $0.30 $1.20 $0.06
minimax/minimax-m2.5 $0.12 $0.99 $0.06
z-ai/glm-5.1 $0.95 $3.15 $0.47
z-ai/glm-5-turbo $1.20 $4.00 $0.24
moonshotai/kimi-k2.5 $0.38 $1.72 $0.19
x-ai/grok-4.20-beta
nvidia/nemotron-3-super-120b-a12b $0.10 $0.50 $0.10
arcee-ai/trinity-large-thinking $0.22 $0.85
openai/gpt-5.4-pro $30.00 $180.00
openai/gpt-5.4-nano $0.20 $1.25 $0.02

── Upgrade at https://portal.nousresearch.com for paid models ──

Available free models:
-> xiaomi/mimo-v2-pro free free free
Enter custom model name
Skip (keep current)

#这里默认会显示小米的免费模型
回车即可

#下一步设置消息通道
Connect a messaging platform? (Telegram, Discord, etc.)

→ Set up messaging now (recommended)
Skip — set up later with 'hermes setup gateway'

#现在默认可选的通道
◆ Messaging Platforms
Connect to messaging platforms to chat with Hermes from anywhere.
Toggle with Space, confirm with Enter.

Select platforms to configure:
Toggle by number, Enter to confirm.

[ ] 1. Telegram
[ ] 2. Discord
[ ] 3. Slack
[ ] 4. Matrix
[ ] 5. Mattermost
[ ] 6. WhatsApp
[ ] 7. Weixin (WeChat)
[ ] 8. BlueBubbles (iMessage)
[ ] 9. Webhooks (GitHub, GitLab, etc.)

─── 💬 Weixin / WeChat Setup ───

1. Hermes will open Tencent iLink QR login in this terminal.
2. Use WeChat to scan and confirm the QR code.
3. Hermes will store the returned account_id/token in ~/.hermes/.env.
4. This adapter supports native text, image, video, and document delivery.

Start QR login now? [Y/n]: y
请使用微信扫描以下二维码：
https://liteapp.weixin.qq.com/q/7GiQu1?qrcode=fd0d4e6ba3320166a66a0adf9f62a391&bot_type=3
（终端二维码渲染失败，请直接打开上面的二维码链接）
会生成二维码链接，打开扫描即可

How should direct messages be authorized?

→ Use DM pairing approval (recommended)
Allow all direct messages
Only allow listed user IDs
Disable direct messages
回车

How should group chats be handled?

→ Disable group chats (recommended)
Allow all group chats
Only allow listed group chat IDs
回车

🚀 Ready to go!

hermes Start chatting
hermes gateway Start messaging gateway
hermes doctor Check for issues

说明安装成功

◆ Messaging Platforms
Connect to messaging platforms to chat with Hermes from anywhere.
Toggle with Space, confirm with Enter.

Select platforms to configure:
Toggle by number, Enter to confirm.

[ ] 1. Telegram
[ ] 2. Discord
[ ] 3. Slack
[ ] 4. Signal
[ ] 5. Email
[ ] 6. SMS (Twilio)
[ ] 7. Matrix
[ ] 8. Mattermost
[ ] 9. WhatsApp
[ ] 10. DingTalk
[ ] 11. Feishu / Lark
[ ] 12. WeCom (Enterprise WeChat)
[ ] 13. WeCom Callback (Self-Built App)
[ ] 14. Weixin (WeChat)
[ ] 15. BlueBubbles (iMessage)
[ ] 16. QQ Bot
[ ] 17. Webhooks (GitHub, GitLab, etc.)

Toggle # (or Enter to confirm):
输入16回车两次

─── 🐧 QQ Bot Setup ───

1. Register a QQ Bot application at q.qq.com
2. Note your App ID and App Secret from the application page
3. Enable the required intents (C2C, Group, Guild messages)
4. Configure sandbox or publish the bot

Your QQ Bot App ID from q.qq.com.
QQ Bot App ID:
输入QQbot appid
✓ Saved QQ_APP_ID

Your QQ Bot App Secret from q.qq.com.
QQ Bot App Secret:
输入QQbot Secret
✓ Saved QQ_CLIENT_SECRET

Optional — restrict DM access to specific user OpenIDs.
The gateway DENIES all users by default for security.
Enter user IDs to create an allowlist, or leave empty
and you'll be asked about open access next.
Allowed user OpenIDs (comma-separated, leave empty for open access):
回车
How should unauthorized users be handled?
↑↓ navigate ENTER/SPACE select ESC cancel
(○) Enable open access (anyone can message the bot)
→ (●) Use DM pairing (unknown users request access, you approve with 'hermes pairing approve')
(○) Skip for now (bot will deny all users until configured)
回车
✓ DM pairing mode — users will receive a code to request access.
Approve with: hermes pairing approve <platform> <code>
OpenID to deliver cron results and notifications to.
Home channel (user/group OpenID for cron delivery, or empty):
回车
✓ 🐧 QQ Bot configured!
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
✓ Messaging platforms configured!
⚠ No home channel set for: QQBot
Without a home channel, cron jobs and cross-platform
messages can't be delivered to those platforms.
Set one later with /set-home in your chat, or:
hermes config set QQBOT_HOME_CHANNEL <channel_id>

Install the gateway as a systemd service? (runs in background, starts on boot) [Y/n]:
输入y回车

Choose how the gateway should run in the background:
↑↓ navigate ENTER/SPACE select ESC cancel

→ (●) User service (no sudo; best for laptops/dev boxes; may need linger after logout)
(○) System service (starts on boot; requires sudo; still runs as your user)
(○) Skip service install for now
回车
Installing user systemd service to: /home/hello/.config/systemd/user/hermes-gateway.service
Created symlink /home/hello/.config/systemd/user/default.target.wants/hermes-gateway.service → /home/hello/.config/systemd/user/hermes-gateway.service.

✓ User service installed and enabled!

Next steps:
hermes gateway start # Start the service
hermes gateway status # Check status
journalctl --user -u hermes-gateway -f # View logs

Enabling linger so the gateway survives SSH logout...
✓ Linger enabled — gateway will persist after logout

Start the service now? [Y/n]:
输入y回车

🚀 Ready to go!

hermes Start chatting
hermes gateway Start messaging gateway
hermes doctor Check for issues

Launch hermes chat now? [Y/n]:
输入y回车默认进入命令行聊天模式 ctrl+c可退出，也可以输入n不进入

hermes gateway install
hermes gateway start

如果微信 发消息没有回复可以尝试

# 查看配对请求
hermes pairing list
# 批准配对请求
hermes pairing approve weixin DLD2Z4DX

或者查看gateway日志

hermes logs gateway

如果出现

WARNING gateway.run: No user allowlists configured. All unauthorized users will be denied. Set GATEWAY_ALLOW_ALL_USERS=true in ~/.hermes/.env to allow open access, or configure platform allowlists (e.g., TELEGRAM_ALLOWED_USERS=your_id).
WARNING gateway.run: Unauthorized user: o9cq80yj2Ad_3UPXjjYoA7wGSBIQ@im.wechat (o9cq80yj2Ad_3UPXjjYoA7wGSBIQ@im.wechat) on weixin
WARNING gateway.run: Unauthorized user: o9cq80yj2Ad_3UPXjjYoA7wGSBIQ@im.wechat (o9cq80yj2Ad_3UPXjjYoA7wGSBIQ@im.wechat) on weixin

可以编辑~/.hermes/.env

WEIXIN_ALLOWED_USERS=o9cq80yj2Ad_3UPXjjYoA7wGSBIQ

之后 微信可以正常发消息使用hermesagent了

git clone https://github.com/nesquena/hermes-webui.git hermes-webui
cd hermes-webui
python3 bootstrap.py

启动后会占用http://127.0.0.1:8787

vi setup-caddy-hermes-agent.sh

内容如下

#!/bin/bash

# ====================== 配置 ======================
LOCAL_PORT="8787" # 本地要代理的端口
PROXY_PORT="80" # 对外暴露的端口（HTTP）

# ====================== 安全停止 caddy ======================
echo "🔌 停止 Caddy 服务..."
systemctl stop caddy 2>/dev/null
sleep 1

# ====================== 判断是否安装 Caddy ======================
if ! command -v caddy &> /dev/null; then
echo "🟢 开始安装 Caddy..."
apt update -y
apt install -y debian-keyring debian-archive-keyring apt-transport-https curl
curl -1sLf 'https://dl.cloudsmith.io/public/caddy/stable/gpg.key' | gpg --dearmor -o /usr/share/keyrings/caddy-stable-archive-keyring.gpg
curl -1sLf 'https://dl.cloudsmith.io/public/caddy/stable/debian.deb.txt' | tee /etc/apt/sources.list.d/caddy-stable.list
apt update -y
apt install -y caddy
fi

# ====================== 写入纯HTTP配置（无证书） ======================
echo "⚙️ 生成 Caddy 配置..."
tee /etc/caddy/Caddyfile > /dev/null <<EOF
:${PROXY_PORT} {
reverse_proxy http://127.0.0.1:${LOCAL_PORT} {
header_up X-Forwarded-Proto http
header_up X-Forward-Host {host}
header_up X-Real-IP {remote_host}
}
header Access-Control-Allow-Origin "*"
}
EOF

# ====================== 启动 ======================
echo "🚀 启动 Caddy..."
systemctl daemon-reload
systemctl enable --now caddy
systemctl restart caddy

echo "=============================================="
echo "✅ 安装配置完成！"
echo "🔗 访问：http://本机IP"
echo "=============================================="

chmod +x setup-caddy-hermes-agent.sh
sudo ./setup-caddy-hermes-agent.sh

此时局域网其他点可用http://ip 访问hermes-webui

Select provider:
↑↓ navigate ENTER/SPACE select ESC cancel

(○) Nous Portal (Nous Research subscription)
(○) OpenRouter (100+ models, pay-per-use)
(○) Anthropic (Claude models — API key or Claude Code)
(○) OpenAI Codex
(○) Xiaomi MiMo (MiMo-V2 models — pro, omni, flash)
(○) Qwen OAuth (reuses local Qwen CLI login)
(○) GitHub Copilot (uses GITHUB_TOKEN or gh auth token)
(○) GitHub Copilot ACP (spawns `copilot --acp --stdio`)
(○) Hugging Face Inference Providers (20+ open models)
(○) Google AI Studio (Gemini models — OpenAI-compatible endpoint)
(○) Google Gemini via OAuth + Code Assist (free tier supported; no API key needed)
(○) DeepSeek (DeepSeek-V3, R1, coder — direct API)
(○) xAI (Grok models — direct API)
(○) Z.AI / GLM (Zhipu AI direct API)
(○) Kimi Coding Plan (api.kimi.com) & Moonshot API
(○) Kimi / Moonshot China (Moonshot CN direct API)
(○) MiniMax (global direct API)
→ (●) MiniMax China (domestic direct API) ← currently active
(○) Alibaba Cloud / DashScope Coding (Qwen + multi-provider)
(○) Ollama Cloud (cloud-hosted open models — ollama.com)
(○) Arcee AI (Trinity models — direct API)
(○) Kilo Code (Kilo Gateway API)
(○) OpenCode Zen (35+ curated models, pay-as-you-go)
(○) OpenCode Go (open models, $10/month subscription)
(○) Vercel AI Gateway (200+ models, pay-per-use)
(○) AWS Bedrock (Claude, Nova, Llama, DeepSeek — IAM or API key)
(○) Custom endpoint (enter URL manually)
(○) Cancel

选择MiniMax (China)

No MiniMax (China) API key configured.
MINIMAX_CN_API_KEY (or Enter to cancel):
#输入apikey
API key saved.

Base URL [https://api.minimaxi.com/anthropic]:
Showing 4 curated models — use "Enter custom model name" for others.
Select default model:
-> MiniMax-M2.7
MiniMax-M2.5
MiniMax-M2.1
MiniMax-M2
Enter custom model name
Skip (keep current)
回车选择默认模型

Hermes agent没有技能市场。