hmo/wechat-hermes-gateway
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

init: WeChat Hermes Gateway - wxhook + Hermes AI auto-reply bot

Browse Source
This commit is contained in:
hmo
2026-05-17 14:13:14 +08:00
commit 93f4273b3d
5 changed files with 564 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.gitignore
+16
View File
@@ -0,0 +1,16 @@
# Python
__pycache__/
*.pyc
.venv/
*.egg-info/
# OS
.DS_Store
Thumbs.db
# IDE
.vscode/
.idea/
# Logs
*.log
README.md
+253
View File
@@ -0,0 +1,253 @@
# 🤖 WeChat Hermes Gateway
Windows 微信机器人 ↔ Linux Hermes AI，全自动双向聊天。
---
## 最终架构
```
┌──────────────────────────────────────────────────────────────┐
│ Windows 192.168.0.111 │
│ │
│ ┌──────────────────┐ ┌───────────────────────────┐ │
│ │ 微信 3.9.5.81 x64 │ │ 日常微信 WeChatAppEx 4.x │ │
│ │ 机器人号modachen │ │ 老爸日常使用，互不干扰 │ │
│ │ wxhook Bot类TCP │ └───────────────────────────┘ │
│ └────────┬─────────┘ │
│ │ wxhook DLL 收消息 │
│ ┌────────▼────────────────┐ │
│ │ wechat_agent.py │ ← 常驻进程 │
│ │ │ │
│ │ Bot类接收微信事件(TCP) │ │
│ │ → POST Hermes API :8642 │ ← 带session: sisyphus │
│ │ ← 收回复 → wxhook发回 │ │
│ │ :5801 ←Hermes找小小莫 │ ← 双向通道 │
│ │ 看门狗防 wxhook 挂掉 │ │
│ └────────┬────────────────┘ │
└───────────┼─────────────────────────────────────────────────┘
│ HTTP (局域网)
┌───────────▼─────────────────────────────────────────────────┐
│ Linux 192.168.0.103 │
│ │
│ ┌──────────────────────────────┐ │
│ │ Hermes Gateway │ │
│ │ - API Server :8642 │ ← OpenAI兼容API │
│ │ - session自动重置: 已关闭 │ ← sisyphus永不清 │
│ │ - 健康检查 /health │ │
│ └──────────────────────────────┘ │
│ │
│ hermes CLI — AI 处理引擎 │
│ 莫荷/莫小荷 — 老爸专属称呼 │
└──────────────────────────────────────────────────────────────┘
```
## 数据流
### 文字消息
```
老爸发微信
→ WeChat 3.9.5.81 收到
→ wxhook DLL TCP → Bot 类 → on_msg 处理器
→ wechat_agent.py POST Hermes API (:8642)
→ X-Hermes-Session-Id: sisyphus (固定)
→ Hermes 处理 → 返回回复
→ wechat_agent.py 收回复 → wxhook API 发回
→ 老爸手机收到
```
### 图片消息
```
老爸发图片
→ WeChat 收到 → wxhook IMAGE_MESSAGE 事件
→ wechat_agent.py 保存图片 → 调豆包OCR (VolcEngine)
→ OCR 文字结果 + 通知 → POST Hermes API
→ Hermes 知道图片内容 → 回复老爸
```
### Hermes 找小小莫（双向）
```
Hermes → POST http://192.168.0.111:5801/hermes-msg
→ wechat_agent.py 写入日志和 inbox 文件
```
## 人物 / ID
| 角色 | 微信名 | wxid | 说明 |
|------|--------|------|------|
| 老爸 | 莫语不语 | `wxid_c0a6izmwd78y22` | 用户，主人 |
| 莫荷/莫小荷 | modachenchen | `wxid_7onnerpx2s2l22` | Hermes AI，老爸专属称呼"莫小荷" |
| 小小莫 | — | — | Sisyphus，Windows 运维，通过 API 与 Hermes 通信 |
## 关键端口
| 端口 | 用途 | 所在 |
|------|------|------|
| 19001 | wxhook HTTP API | Windows |
| 5801 | Hermes→小小莫 消息入口 | Windows |
| 8642 | Hermes API Server (OpenAI兼容) | Linux |
| 5800 | bridge.py (旧，不再使用) | Linux |
## 组件
### Windows 端（wechat_agent.py）
- **wxhook Bot 类** — DLL 注入 + TCP 收消息
- **Hermes API 调用** — 直接 POST :8642，session 固定 `sisyphus`
- **回复服务** — 5801 端口收 Hermes 消息
- **看门狗** — 2 分钟无消息自动刷新 webhook；API 挂了才重注入 DLL
- **昵称缓存** — 从 wxhook getContactList 获取联系人昵称
### Linux 端（Hermes Gateway）
- **API Server** — 0.0.0.0:8642，Bearer auth
- **session 管理** — `api_server` 平台关闭自动重置，`sisyphus` 永不清上下文
- **配置位置** — `/home/hmo/.hermes/config.yaml`
- **Provider** — `ocg-new` → `https://opencode.ai/zen/go/v1`
## 启动步骤
### Windows
```batch
cd D:\F\NewI\opencode\daily-workspace\projects\wechat-hermes-gateway
scripts\start_bridge.bat
```
或直接：
```powershell
$env:PYTHONHOME=''
Start-Process -WindowStyle Hidden python.exe scripts\wechat_agent.py
```
启动后需用修复过低工具扫码登录微信。
### Linux（如重启后）
```bash
source /home/hmo/hermes-agent/.venv/bin/activate
hermes gateway restart
```
验证：
```bash
ss -tlnp | grep 8642
curl http://127.0.0.1:8642/v1/models
```
## 通信方式
| 方向 | 方式 | 示例 |
|------|------|------|
| 小小莫 → Hermes | POST :8642/v1/chat/completions | 带 `X-Hermes-Session-Id: sisyphus` |
| Hermes → 小小莫 | POST :5801/hermes-msg | 写入 `C:\Users\hmo\Desktop\hermes_inbox.txt` |
| 老爸 ↔ Hermes | 微信聊天 | 自动通过 wechat_agent.py 桥接 |
## 项目文件
```
wechat-hermes-gateway/
├── README.md # 本文档
├── scripts/
│ ├── wechat_agent.py # 主力：微信机器人代理
│ └── start_bridge.bat # 一键启动脚本
└── temp/ # 废弃/临时脚本
```
## 历史决策
1. **wxhook HTTP webhook 不可靠** → 改用 Bot 类 TCP 收消息
2. **Bot 类偶尔停发事件** → 加看门狗自动刷新
3. `hermes -z` **无上下文** → 改用 Hermes API Server (:8642) + session
4. **session 自动重置** → 关闭 api_server 平台的重置策略
5. **群聊不认人** → session 固定 `sisyphus`，所有消息共享上下文
6. **Linux bridge 常挂** → 去掉 bridge.py，Windows 直接调 Hermes API
## 已实现的功能
| 功能 | 状态 |
|------|------|
| 文字消息收发（个人聊天） | ✅ 双向，session 上下文连贯 |
| 文字消息收发（群聊） | ✅ 同 session，认识老爸 |
| 图片消息接收 + OCR 分析 | ✅ 自动保存 → 豆包 OCR → 结果给 Hermes |
| Hermes 身份认知 | ✅ 知道自己是莫荷/莫小荷，知道老爸 |
| 会话上下文持续 | ✅ session `sisyphus`，自动重置已关闭 |
| 小小莫 ↔ Hermes 双向通信 | ✅ API (:8642) + HTTP (:5801/hermes-msg) |
| wxhook 看门狗自愈 | ✅ 2分钟无消息刷新 webhook |
| 昵称识别 | ✅ 从 getContactList 获取 |
| 联系人列表查询 | ✅ wxhook /api/getContactList |
## 未实现 / 不可行
| 功能 | 原因 |
|------|------|
| 语音消息 | ⏳ 已能检测并下载，转文字（STT）待接入 |
| 发送图片 | wxhook 有 send_image API 但未接入回复链路 |
| 文件收发 | 同上，未接入 |
| 换头像/改资料 | wxhook 无相关 API |
| 群管理（拉人踢人） | wxhook 无群管理 API |
| iLink 官方 bot 接口 | 限制太多，弃用 |
| 多人独立会话 | 目前全部共享 `sisyphus` 单会话 |
## 灾难恢复流程
### 场景：Windows 重启
需要 3 步，**顺序不能错**：
```
第1步：双击 start_bridge.bat
→ 自动杀旧微信 → Bot 类启动新微信 + 注入 DLL
→ 等待微信窗口出现
第2步：运行修复过低工具
→ 选择修复过低6.0\低版通用杀器.sp.exe
→ 自动扫描到已运行的微信 → 打补丁 → 弹出登录二维码
第3步：手机扫码登录
→ 登录后 wechat_agent.py 自动检测到登录状态
→ 开始转发消息
→ 给 filehelper 发 "[Bridge] online" 确认
```
验证：手机发条消息给 modachenchen，看能否收到自动回复。
### 场景：Linux 重启
```bash
# 1. 启动 Hermes gateway（自动恢复 session）
source /home/hmo/hermes-agent/.venv/bin/activate
hermes gateway restart
# 2. 验证
ss -tlnp | grep 8642 # 确认 API 端口
curl http://127.0.0.1:8642/v1/models # 确认 API 响应
# 3. 确认 Windows 能连上
# 从 Windows 运行：
curl http://192.168.0.103:8642/v1/models -H "Authorization: Bearer hermes123"
```
### 场景：两边都重启了
1. Linux 先：`hermes gateway restart` + 验证 8642 监听
2. Windows 后：`start_bridge.bat` → 修复工具登录 → 完成
### 场景：Hermes 不认人了（session 丢了）
不用慌，system prompt 里已经写死了她的身份和你的身份。
发条消息她就会看到：
> "你是莫荷，女生。你的主人是老爸（微信名：莫语不语）"
如果连这都不奏效 → 告诉 Hermes "去找小小莫" → 它会 POST 到 :5801/hermes-msg → 我来处理。
## 注意事项
- wxhook DLL 仅支持 x64 微信 3.9.5.81
- 每次 WeChat 重启需重新登录（修复过低工具）
- **start_bridge.bat 必须第 1 步执行**，修复工具第 2 步
- Hermes API 首次调用可能较慢（大模型冷启动）
- 看门狗只刷新 webhook，不会误伤正常消息处理
- 如果微信登录后没反应，等 1-2 分钟看门狗会自动刷新
scripts/bridge.py
+74
View File
@@ -0,0 +1,74 @@
"""
WeChat Hermes Bridge — with webhook keepalive
"""
import pymem, pymem.process, requests, time, json, threading
from http.server import HTTPServer, BaseHTTPRequestHandler
DLL = r"C:\Users\hmo\AppData\Local\Programs\Python\Python310\Lib\site-packages\wxhook\tools\wxhook.dll"
WX = "http://127.0.0.1:19088"
HERMES = "http://192.168.0.103:5800/callback"
LOG = r"C:\Users\hmo\Desktop\bridge.log"
BOT = "wxid_7onnerpx2s2l22"
PORT = 5801
def log(m):
with open(LOG, "a", encoding="utf-8") as f:
f.write(f"{time.strftime('%H:%M:%S')} {m}\n")
# Inject DLL
try:
requests.post(f"{WX}/api/checkLogin", json={}, timeout=3)
log("DLL OK")
except:
pm = pymem.Pymem("WeChat.exe")
pymem.process.inject_dll(pm.process_handle, DLL.encode())
time.sleep(2)
log("DLL injected")
log(f"login: {requests.post(f'{WX}/api/checkLogin', json={}, timeout=5).json()}")
# Periodic webhook refresh (every 30 seconds)
def webhook_keepalive():
while True:
try:
requests.post(f"{WX}/api/hookSyncMsg", json={
"port": PORT, "ip": "0.0.0.0", "enableHttp": 1,
"url": f"http://127.0.0.1:{PORT}", "timeout": 300
}, timeout=5)
except:
pass
time.sleep(30)
threading.Thread(target=webhook_keepalive, daemon=True).start()
log("keepalive started (30s)")
# Test message
requests.post(f"{WX}/api/sendTextMsg",
json={"wxid": "filehelper", "msg": "[Bridge] online"}, timeout=5)
class H(BaseHTTPRequestHandler):
def do_POST(self):
body = self.rfile.read(int(self.headers.get("Content-Length", 0)))
try:
d = json.loads(body)
fu = d.get("fromUser", "")
ct = d.get("content", "")
if fu and ct and fu != BOT:
log(f"MSG {fu}: {ct[:60]}")
threading.Thread(target=lambda: requests.post(HERMES, json=[{"id": int(time.time()*1000), "type": 1, "content": ct, "sender": fu, "roomid": "", "ts": time.time()}], timeout=30), daemon=True).start()
self.send_response(200); self.end_headers(); return
to = d.get("to", "") or d.get("wxid", "")
msg = d.get("message", "") or d.get("content", "")
if to and msg:
r = requests.post(f"{WX}/api/sendTextMsg", json={"wxid": to, "msg": msg}, timeout=5)
log(f"SEND {to}: {r.get('msg','')}")
self.send_response(200); self.end_headers(); return
except Exception as e:
log(f"ERR: {str(e)[:80]}")
self.send_response(200); self.end_headers()
def do_GET(self):
self.send_response(200); self.end_headers(); self.wfile.write(b'{"ok":true}')
def log_message(self, *a): pass
log(f"ready on :{PORT}")
HTTPServer(("0.0.0.0", PORT), H).serve_forever()
scripts/start_bridge.bat
+34
View File
@@ -0,0 +1,34 @@
@echo off
title WeChat Hermes Bridge
cd /d "%~dp0.."
set PYTHON=C:\Users\hmo\AppData\Local\Programs\Python\Python310\python.exe
set AGENT=scripts\wechat_agent.py
echo ========================================
echo WeChat Hermes Bridge
echo ========================================
echo.
echo [1/3] 启动微信 + 注入 DLL...
echo.
set PYTHONHOME=
set WXHOOK_LOG_LEVEL=ERROR
%PYTHON% %AGENT%
echo.
echo ========================================
echo 微信已启动！
echo.
echo 第 2 步：运行修复过低工具 ^> 扫码登录
echo 路径：D:\Program Files (x86)\低版本修复工具\
echo 低版本修复工具\修复过低6.0\低版通用杀器.sp.exe
echo.
echo 第 3 步：手机扫码 → 自动开始转发
echo.
echo 按 Ctrl+C 停止桥接
echo ========================================
if %errorlevel% neq 0 (
pause
)
scripts/wechat_agent.py
+187
View File
@@ -0,0 +1,187 @@
"""
WeChat Agent - wxhook + Hermes API (:8642)
"""
import sys, os, json, time, threading, requests, re
from http.server import HTTPServer, BaseHTTPRequestHandler
sys.path.insert(0, r"C:\Users\hmo\AppData\Local\Programs\Python\Python310\Lib\site-packages")
os.environ["WXHOOK_LOG_LEVEL"] = "ERROR"
from wxhook import Bot
from wxhook.events import TEXT_MESSAGE, IMAGE_MESSAGE, VOICE_MESSAGE
import pymem, pymem.process
BOT_WXID = "wxid_7onnerpx2s2l22"
WX_API = ""
LOG_FILE = r"C:\Users\hmo\Desktop\wechat_agent.log"
DLL = r"C:\Users\hmo\AppData\Local\Programs\Python\Python310\Lib\site-packages\wxhook\tools\wxhook.dll"
last_msg_time = time.time()
nickname_cache = {}
HERMES_API = "http://192.168.0.103:8642/v1/chat/completions"
HERMES_KEY = "hermes123"
def log(m):
with open(LOG_FILE, "a", encoding="utf-8") as f:
f.write(f"{time.strftime('%H:%M:%S')} {m}\n")
def send_wx(wxid, msg):
try:
requests.post(WX_API + "/api/sendTextMsg", json={"wxid": wxid, "msg": msg}, timeout=5)
log(f"SEND {wxid}: ok")
except Exception as e:
log(f"SEND ERR: {e}")
def get_nickname(wxid):
if wxid in nickname_cache:
return nickname_cache[wxid]
try:
r = requests.post(WX_API + "/api/getContactList", json={}, timeout=5)
for c in r.json().get("data", []):
if c.get("wxid") == wxid:
nick = c.get("nickname") or c.get("customAccount") or wxid
nickname_cache[wxid] = nick
return nick
except:
pass
nickname_cache[wxid] = wxid
return wxid
def call_hermes(wxid, content):
nickname = get_nickname(wxid)
headers = {"Authorization": f"Bearer {HERMES_KEY}", "X-Hermes-Session-Id": "sisyphus", "Content-Type": "application/json"}
sys_prompt = f"你是莫荷，女生。你主人是老爸（{nickname}）。回复简短像聊天。发图用 [IMG]URL[/IMG]。"
body = {"model": "hermes-agent", "messages": [{"role": "system", "content": sys_prompt}, {"role": "user", "content": content}]}
try:
r = requests.post(HERMES_API, json=body, headers=headers, timeout=120, proxies={"http": None, "https": None})
if r.status_code == 200:
return r.json()["choices"][0]["message"]["content"]
except Exception as e:
log(f"API ERR: {e}")
return None
def watchdog():
global last_msg_time
while True:
idle = time.time() - last_msg_time
if idle > 120 and WX_API:
try:
r = requests.post(WX_API + "/api/checkLogin", json={}, timeout=5)
if r.json().get("code") == 1:
requests.post(WX_API + "/api/hookSyncMsg", json={"ip": "127.0.0.1", "port": 19001, "enableHttp": 1, "url": "", "timeout": 300}, timeout=5)
log(f"WATCHDOG: refreshed ({int(idle)}s)")
else:
log("WATCHDOG: re-injecting...")
pymem.process.inject_dll(pymem.Pymem("WeChat.exe").process_handle, DLL.encode())
except:
pass
last_msg_time = time.time()
time.sleep(30)
threading.Thread(target=watchdog, daemon=True).start()
class RH(BaseHTTPRequestHandler):
def do_POST(self):
global last_msg_time
last_msg_time = time.time()
body = self.rfile.read(int(self.headers.get("Content-Length", 0)))
try:
d = json.loads(body)
if self.path == "/hermes-msg":
msg = d.get("message", "") or d.get("content", "") or str(d)[:200]
log("<<< HERMES: " + msg[:100])
with open(r"C:\Users\hmo\Desktop\hermes_inbox.txt", "a", encoding="utf-8") as f:
f.write(f"{time.strftime('%H:%M:%S')} {msg}\n")
self.send_response(200); self.end_headers(); return
to = d.get("to", "") or d.get("wxid", "")
msg = d.get("message", "") or d.get("content", "")
if to and msg:
log(f"REPLY {to}: {msg[:50]}")
send_wx(to, msg)
except Exception as e:
log(f"RH ERR: {e}")
self.send_response(200); self.end_headers()
def do_GET(self):
self.send_response(200); self.end_headers(); self.wfile.write(b'{"ok":true}')
def log_message(self, *a): pass
threading.Thread(target=lambda: HTTPServer(("0.0.0.0", 5801), RH).serve_forever(), daemon=True).start()
log("HTTP :5801")
log("Creating Bot...")
b = Bot()
WX_API = b.BASE_URL
log("Bot ready, API=" + WX_API)
@b.handle([TEXT_MESSAGE, IMAGE_MESSAGE, VOICE_MESSAGE])
def on_msg(_bot, event):
global last_msg_time
last_msg_time = time.time()
fu = event.fromUser or ""
if not fu or fu == BOT_WXID:
return
if event.type == VOICE_MESSAGE:
log(f"<- {fu}: [voice]")
reply = call_hermes(fu, "[voice message]")
if reply: send_wx(fu, reply)
return
if event.type == IMAGE_MESSAGE:
log(f"<- {fu}: [image]")
b64 = event.base64Img or ""
ocr_text = ""
if b64:
try:
import base64
os.makedirs(r"C:\Users\hmo\Desktop\wechat_images", exist_ok=True)
fname = os.path.join(r"C:\Users\hmo\Desktop\wechat_images", str(int(time.time())) + ".jpg")
with open(fname, "wb") as f:
f.write(base64.b64decode(b64))
api_key = os.environ.get("VOLCENGINE_API_KEY", "b0359bed-09f2-49e2-a53c-32ba057412e3")
with open(fname, "rb") as f:
img_b64 = base64.b64encode(f.read()).decode()
r = requests.post("https://ark.cn-beijing.volces.com/api/coding/v3/chat/completions",
json={"model": "doubao-seed-code", "messages": [{"role": "user", "content": [{"type": "text", "text": "描述图片"}, {"type": "image_url", "image_url": {"url": "data:image/jpeg;base64," + img_b64}}]}], "max_tokens": 500},
headers={"Authorization": "Bearer " + api_key, "Content-Type": "application/json"}, timeout=60)
ocr_text = r.json()["choices"][0]["message"]["content"]
except Exception as e:
log(f"OCR err: {e}")
msg = "[image]" + ("\n" + ocr_text if ocr_text else "")
reply = call_hermes(fu, msg)
if reply: send_wx(fu, reply)
return
content = event.content or ""
if not content:
return
log(f"<- {fu}: {content[:50]}")
reply = call_hermes(fu, content)
if reply:
log(f"-> {fu}: {reply[:50]}")
img_match = re.search(r'\[IMG\](.*?)\[/IMG\]', reply)
if img_match:
img_cmd = img_match.group(1).strip()
clean = re.sub(r'\s*\[IMG\].*?\[/IMG\]\s*', '', reply).strip()
if clean:
send_wx(fu, clean)
try:
ir = requests.get(img_cmd, timeout=30, proxies={"http": None, "https": None})
if ir.status_code == 200:
ext = ".jpg"
if "png" in ir.headers.get("content-type", ""): ext = ".png"
tmp = os.path.join(r"C:\Users\hmo\Desktop", f"send_img_{int(time.time())}{ext}")
with open(tmp, "wb") as f:
f.write(ir.content)
try:
_bot.send_image(fu, tmp)
except:
requests.post(WX_API + "/api/sendImagesMsg", json={"wxid": fu, "imagePath": tmp}, timeout=10)
os.remove(tmp)
except Exception as e:
log(f"IMG err: {e}")
else:
send_wx(fu, reply)
else:
log(f"-> {fu}: no reply")
print("[Agent] Hermes API: " + HERMES_API)
log("Ready")
b.run()