在基于 Debian/Ubuntu 等 Linux 发行版上安装 dae 大鹅裸核代理

下载和安装 dae 大鹅

去 Github 下载对应架构的安装包，我的是 Debian 系统，下载 dae-linux-x86_64.deb；使用以下命令安装：

sudo dpkg -i dae-linux-x86_64.deb

sudo apt install dae-linux-x86_64.deb

以上两个命令随便选一个就行。

dae 大鹅配置文件

编辑或者创建 /etc/dae/config.dae

vi /etc/dae/config.dae

复制粘贴以下内容

# You can split the configuration into multiple files with `include { ... }`.
# include {}

global {
    ##### Software options.
    tproxy_port: 12345
    tproxy_port_protect: true
    pprof_port: 0
    so_mark_from_dae: 0
    log_level: info
    disable_waiting_network: false

    ##### Interface and kernel options.
    # lan_interface: eth0
    wan_interface: auto
    auto_config_kernel_parameter: true

    ##### Node connectivity check.
    tcp_check_url: 'http://www.google.com/gen_204,216.58.200.36,2404:6800:4005:809::2004'
    tcp_check_http_method: GET
    udp_check_dns: 'dns.google:53,8.8.8.8,2001:4860:4860::8888'
    check_interval: 30s
    check_tolerance: 0ms

    ##### Connecting options.
    dial_mode: domain
    allow_insecure: true
    sniffing_timeout: 100ms

    # TLS & uTLS 客户端特征伪装
    tls_implementation: utls
    utls_imitate: chrome_auto
    tls_fragment: false

    # 带宽塑形
    bandwidth_max_tx: '200 mbps'
    bandwidth_max_rx: '1 gbps'
}

subscription {
    my_sub: 'https-file://example.com/sub/url'
}

# 节点配置
node {
    local_socks5: 'socks5://127.0.0.1:1080'
    my_vless: 'vless://auto%uuid@ip:端口?encryption=none&security=tls&sni=域名&fp=edge#🇳🇱 荷兰 re'
    my_vmess: 'vmess://uuid@域名:端口?encryption=auto&type=ws&security=tls&host=de010.域名&path=%2Fprofile%2Fde09#🇩🇪 德国 10'
}

# 分组配置
group {
    proxy {
        # 🇺🇸 节点选择策略
        policy: min_moving_avg

        # 指定机场节点
        filter: subtag(my_sub) && name(keyword: '美国 01')
    }
# 💬 社媒专用分组
    social_proxy {
        # 可根据需要选择 min_moving_avg 或 latency
        policy: min_moving_avg

        # 自建家庭宽带线路适合流媒体/社媒的节点
        filter: name(keyword: '德国 10')
    }
}

# DNS 配置：upstream 与 DNS 分流规则必须嵌套在 dns{} 内，
dns {
    upstream {
        cnnicdns: 'tcp+udp://1.8.8.8:53'
        onedns: 'tcp+udp://52.80.66.66:53'
        localdns: 'tcp+udp://192.168.1.1:53'
        googledns: 'h3://8.8.8.8/dns-query'
        securedns: 'https://doh.360.cn/dns-query'
        cloudflare-dns: 'h3://1.1.1.1/dns-query'
    }

    routing {
        request {
            qname(geosite:adobe, geosite:linkedin, geosite:meta, geosite:github, geosite:google, geosite:x) -> cloudflare-dns
            qname(geosite:cn, geosite:alibaba, geosite:debian, geosite:category-ru) -> localdns
            qname(geosite:adguard, geosite:amazon, geosite:apple, geosite:microsoft, geosite:mozilla) -> securedns
            qname(geosite:lark, geosite:category-games, geosite:category-dev) -> cnnicdns
            qname(suffix:bjxuejing.cn, suffix: meixi-mgo.com, suffix: xzmgo.com, suffix:tuyoo.com, suffix:wanyiwan.top) -> onedns
        fallback: googledns
        }
    }
}

# 路由匹配核心
routing {
    ### Preset rules.
    pname(NetworkManager, systemd-resolved, dnsmasq, gnome-weather, sing-box) -> must_direct
    dip(224.0.0.0/3, 'ff00::/8') -> direct
    dip(geoip:private) -> direct

    ### 自定义规则
    domain(suffix:cfysoft.cc, suffix:cndhl.com, suffix:huohua.icu, suffix:jisuimage.com, suffix:mytvapp.cc, suffix:scdn.co, suffix:xmfyy.com) -> direct
    domain(suffix:cdnbuild.net, suffix:cdngslb.com, suffix:ibytedapm.com, suffix:icoa.cn, suffix:icoa.vip, suffix:tigress.cc) -> direct
    domain(suffix:bjxuejing.cn, suffix:edu.kg, suffix:hdcdn.online, suffix:ip.sb, suffix:open-meteo.com, suffix:wanyiwan.top) -> direct
    domain(suffix:beacondb.net, suffix:dmcdn.net, suffix:fuckcloudnative.io, suffix:localsend.org, suffix:oceansaver.in, suffix:rustdesk.com) -> direct
    domain(suffix:dynadot.com, suffix: frantech.ca, suffix: gnome.org, suffix: gwdg.de, suffix: meixi-mgo.com, suffix: xzmgo.com) -> direct
    domain(suffix: accuweather.com, suffix:aviationweather.gov, suffix:met.no, suffix:openweathermap.org, suffix:qweather.com, suffix:weatherapi.com) -> direct
    domain(suffix:analytics.google.com, suffix:ads.google.com, suffix:tpc.googlesyndication.com) -> proxy
    domain(suffix:addons.mozilla.org, suffix:onedrive.live.com, suffix:primevideo.com) -> proxy
    domain(suffix:flathub.org, suffix:ml0987.com, suffix:openwrt.org, suffix:steamcommunity.com) -> proxy
    domain(geosite:category-ads) -> block
    dip(geoip:cn, geoip:private, 8.219.1.15, 47.245.124.74) -> direct
    dip(91.108.4.0/22, 91.108.8.0/22, 91.108.12.0/22, 91.108.16.0/22, 91.108.20.0/22, 91.108.36.0/23, 91.108.38.0/23, 91.108.56.0/22) -> proxy
    dip(149.154.160.0/20, 149.154.164.0/22, 149.154.168.0/22, 149.154.172.0/22, 185.76.192.0/22) -> proxy
    dip('2001:b28:f23c::/47', '2001:b28:f23f::/48', '2001:67c:4e8::/48') -> proxy
    domain(geosite:github, geosite:google, geosite:linkedin, geosite:meta, geosite:paypal, geosite:telegram, geosite:tiktok, geosite:x) -> proxy
    domain(geosite:alibaba, geosite:cn, geosite:adobe, geosite:category-ru) -> direct
    domain(geosite:amazon, geosite:apple, geosite:microsoft, geosite:mozilla) -> direct
    domain(geosite:category-games, geosite:category-dev, geosite:category-browser-!cn) -> direct
    fallback: proxy
}

如果你是软路由，应该取消注释 # lan_interface: eth0，并使用以下命令查看所有网卡及其 IP 地址

ip addr

或者

ip route show default

WAN 口：通常连接光猫或上级路由器，获取的是运营商的公网 IP/运营商内网 IP，或者是通过 PPPoE 拨号的网卡。

LAN 口：通常连接 ap 或下级设备，上面配置的是你本地局域网的网关 IP（例如 192.168.x.1）。我这里是 Debian 桌面系统，所以只要设置 wan 口 就可以了。

配置里的机场订阅部分，把引号里的内容换成你的订阅链接。’https’ 后加上 ‘-file’，这样在启动 dae 服务后，会向 /etc/dae/persist.d/my_sub.sub 这个文件写入节点信息。

subscription {
    my_sub: 'https-file://example.com/sub/url'
}

检查一下配置信息是否有错

dae validate -c /etc/dae/config.dae

没有错误不会有任何输出

启动 dae 服务，开启科学上网之旅

启动 dae 服务

systemctl start dae

查看 dae 服务状态

systemctl status dae

重启和停止 dae 服务

systemctl restart dae

systemctl stop dae

选择特定节点

要更改节点，首先需要解码 dae 下载的 Base64 订阅，创建以下脚本

vi /etc/dae/convert-sub-view.sh

复制粘贴以下内容

#!/bin/sh

# ============================================================
# dae 订阅可读版生成脚本
#
# 输入：
#   /etc/dae/persist.d/my_sub.sub
#
# 输出：
#   /etc/dae/cnsub.sub
#
# 功能：
#   1. 解码 dae 下载的 Base64 订阅
#   2. 自动识别多种代理协议
#   3. VMess 从 JSON 的 ps 字段读取节点名称
#   4. VLESS / AnyTLS / Trojan / SS / SSR / Hysteria2 /
#      TUIC / Juicity 等从 URI #fragment 读取节点名称
#   5. URL 解码中文、emoji 等节点名称
#   6. 保留原始节点 URI，不修改节点参数
#
# 注意：
#   cnsub.sub 仅供人工查看
#   不加入 config.dae
#   不会影响 dae 运行
# ============================================================

SRC="/etc/dae/persist.d/my_sub.sub"
DST="/etc/dae/cnsub.sub"

if [ ! -f "$SRC" ]; then
    echo "错误：找不到源文件："
    echo "  $SRC"
    exit 1
fi

python3 - "$SRC" "$DST" <<'PY'
import sys
import base64
import json
import urllib.parse
import re
from pathlib import Path


SRC = Path(sys.argv[1])
DST = Path(sys.argv[2])


# ============================================================
# Base64 解码
# ============================================================

def b64decode(data):
    """
    自动处理：
    - 普通 Base64
    - URL-safe Base64
    - 缺少 =
    - 换行/空白
    """

    data = data.strip()
    data = re.sub(r"\s+", "", data)

    # 自动补齐 Base64 padding
    data += "=" * (-len(data) % 4)

    try:
        return base64.urlsafe_b64decode(data)
    except Exception:
        return base64.b64decode(data)


# ============================================================
# URL fragment 解码
# ============================================================

def decode_fragment(fragment):
    """
    将：
        %F0%9F%87%BA%F0%9F%87%B8%20%E7%BE%8E%E5%9B%BD

    转换成：
        🇺🇸 美国
    """

    try:
        return urllib.parse.unquote(fragment)
    except Exception:
        return fragment


# ============================================================
# 判断是否像代理 URI
# ============================================================

PROTOCOLS = (
    "vmess://",
    "vless://",
    "anytls://",
    "trojan://",
    "ss://",
    "ssr://",
    "hysteria://",
    "hysteria2://",
    "hy2://",
    "tuic://",
    "juicity://",
    "socks://",
    "socks5://",
)


def is_proxy_uri(line):
    line_lower = line.lower()

    return line_lower.startswith(PROTOCOLS)


# ============================================================
# 获取 URI fragment 名称
# ============================================================

def get_uri_name(uri):
    """
    从：

        anytls://[email protected]:443/?insecure=1#%E7%BE%8E%E5%9B%BD

    获取：

        美国
    """

    try:
        if "#" not in uri:
            return ""

        fragment = uri.split("#", 1)[1]

        return decode_fragment(fragment).strip()

    except Exception:
        return ""


# ============================================================
# VMess 节点解析
# ============================================================

def get_vmess_info(uri):
    """
    VMess 格式：

        vmess://BASE64(JSON)

    返回：

        节点名称
        JSON 对象
    """

    try:
        payload = uri[len("vmess://"):].strip()

        raw = b64decode(payload)

        obj = json.loads(
            raw.decode(
                "utf-8",
                errors="replace"
            )
        )

        name = obj.get("ps", "")

        if name is None:
            name = ""

        name = str(name).strip()

        return name, obj

    except Exception as e:
        return "", None


# ============================================================
# 解析一个节点
# ============================================================

def parse_node(uri):

    uri = uri.strip()

    # --------------------------------------------------------
    # VMess
    # --------------------------------------------------------

    if uri.lower().startswith("vmess://"):

        name, obj = get_vmess_info(uri)

        if name:
            return "VMess", name

        # 如果 VMess 内部没有 ps
        fragment_name = get_uri_name(uri)

        if fragment_name:
            return "VMess", fragment_name

        return "VMess", "未命名节点"


    # --------------------------------------------------------
    # VLESS
    # --------------------------------------------------------

    if uri.lower().startswith("vless://"):
        name = get_uri_name(uri)

        return "VLESS", name or "未命名节点"


    # --------------------------------------------------------
    # AnyTLS
    # --------------------------------------------------------

    if uri.lower().startswith("anytls://"):
        name = get_uri_name(uri)

        return "AnyTLS", name or "未命名节点"


    # --------------------------------------------------------
    # Trojan
    # --------------------------------------------------------

    if uri.lower().startswith("trojan://"):
        name = get_uri_name(uri)

        return "Trojan", name or "未命名节点"


    # --------------------------------------------------------
    # Shadowsocks
    # --------------------------------------------------------

    if uri.lower().startswith("ss://"):
        name = get_uri_name(uri)

        return "Shadowsocks", name or "未命名节点"


    # --------------------------------------------------------
    # ShadowsocksR
    # --------------------------------------------------------

    if uri.lower().startswith("ssr://"):
        name = get_uri_name(uri)

        return "ShadowsocksR", name or "未命名节点"


    # --------------------------------------------------------
    # Hysteria
    # --------------------------------------------------------

    if uri.lower().startswith("hysteria://"):
        name = get_uri_name(uri)

        return "Hysteria", name or "未命名节点"


    # --------------------------------------------------------
    # Hysteria2
    # --------------------------------------------------------

    if uri.lower().startswith("hysteria2://"):
        name = get_uri_name(uri)

        return "Hysteria2", name or "未命名节点"


    # --------------------------------------------------------
    # HY2
    # --------------------------------------------------------

    if uri.lower().startswith("hy2://"):
        name = get_uri_name(uri)

        return "Hysteria2", name or "未命名节点"


    # --------------------------------------------------------
    # TUIC
    # --------------------------------------------------------

    if uri.lower().startswith("tuic://"):
        name = get_uri_name(uri)

        return "TUIC", name or "未命名节点"


    # --------------------------------------------------------
    # Juicity
    # --------------------------------------------------------

    if uri.lower().startswith("juicity://"):
        name = get_uri_name(uri)

        return "Juicity", name or "未命名节点"


    # --------------------------------------------------------
    # SOCKS
    # --------------------------------------------------------

    if uri.lower().startswith(("socks://", "socks5://")):
        name = get_uri_name(uri)

        return "SOCKS", name or "未命名节点"


    # --------------------------------------------------------
    # 未知协议
    # --------------------------------------------------------

    return "Unknown", get_uri_name(uri) or "未命名节点"


# ============================================================
# 读取原始文件
# ============================================================

raw = SRC.read_text(
    encoding="utf-8",
    errors="replace"
).strip()


if not raw:
    print("错误：源文件为空")
    sys.exit(1)


# ============================================================
# 解码最外层 Base64
# ============================================================

def decode_subscription(data):

    # 如果已经是明文 URI 列表
    if any(
        data.startswith(protocol)
        for protocol in PROTOCOLS
    ):
        return data

    try:

        decoded = b64decode(data).decode(
            "utf-8",
            errors="replace"
        )

        # 判断解码后是否确实像节点列表
        if any(
            line.strip().lower().startswith(PROTOCOLS)
            for line in decoded.splitlines()
        ):
            return decoded

        # 有些订阅可能多套 Base64
        # 再尝试一次
        try:

            decoded2 = b64decode(decoded).decode(
                "utf-8",
                errors="replace"
            )

            if any(
                line.strip().lower().startswith(PROTOCOLS)
                for line in decoded2.splitlines()
            ):
                return decoded2

        except Exception:
            pass

        return decoded

    except Exception:
        return data


decoded = decode_subscription(raw)


# ============================================================
# 逐行解析
# ============================================================

nodes = []

status_lines = []
unknown_lines = []

for line in decoded.splitlines():

    line = line.strip()

    if not line:
        continue


    # --------------------------------------------------------
    # STATUS
    # --------------------------------------------------------

    if line.startswith("STATUS="):

        status_lines.append(line)

        continue


    # --------------------------------------------------------
    # 标准代理 URI
    # --------------------------------------------------------

    if is_proxy_uri(line):

        protocol, name = parse_node(line)

        nodes.append({
            "protocol": protocol,
            "name": name,
            "uri": line
        })

        continue


    # --------------------------------------------------------
    # 其他未知内容
    # --------------------------------------------------------

    unknown_lines.append(line)


# ============================================================
# 生成 cnsub.sub
# ============================================================

with DST.open(
    "w",
    encoding="utf-8"
) as f:

    f.write("# ============================================================\n")
    f.write("# dae 订阅人类可读版\n")
    f.write("# ============================================================\n")
    f.write("#\n")
    f.write("# 本文件仅供人工查看。\n")
    f.write("# 不参与 dae 运行。\n")
    f.write("# 不要将本文件加入 config.dae 的 subscription。\n")
    f.write("#\n")
    f.write(f"# 原始文件：{SRC}\n")
    f.write("# ============================================================\n\n")


    # --------------------------------------------------------
    # STATUS
    # --------------------------------------------------------

    if status_lines:

        f.write("# -------------------- 订阅状态 --------------------\n")

        for status in status_lines:
            f.write(f"# {status}\n")

        f.write("\n")


    # --------------------------------------------------------
    # 节点
    # --------------------------------------------------------

    f.write("# -------------------- 节点列表 --------------------\n\n")


    for index, node in enumerate(nodes, 1):

        f.write(
            f"# {index:03d}. "
            f"[{node['protocol']}] "
            f"{node['name']}\n"
        )

        f.write(node["uri"])
        f.write("\n\n")


    # --------------------------------------------------------
    # 未知内容
    # --------------------------------------------------------

    if unknown_lines:

        f.write(
            "# -------------------- 未识别内容 --------------------\n"
        )

        for line in unknown_lines:
            f.write(f"# {line}\n")

        f.write("\n")


# ============================================================
# 输出统计
# ============================================================

print()
print("============================================================")
print(" dae 订阅人类可读版生成完成")
print("============================================================")
print(f"源文件：{SRC}")
print(f"输出文件：{DST}")
print()
print(f"节点总数：{len(nodes)}")
print()

# 按协议统计
protocol_count = {}

for node in nodes:
    protocol = node["protocol"]

    protocol_count[protocol] = (
        protocol_count.get(protocol, 0) + 1
    )

for protocol, count in protocol_count.items():

    print(
        f"  {protocol:<15} {count:>4}"
    )

print()
print("节点列表：")
print("------------------------------------------------------------")

for index, node in enumerate(nodes, 1):

    print(
        f"{index:03d}. "
        f"[{node['protocol']}] "
        f"{node['name']}"
    )

print("------------------------------------------------------------")
print()
print(f"已生成：{DST}")
print()
PY

exit $?

更改一下脚本权限

chmod +x /etc/dae/convert-sub-view.sh

运行脚本，脚本会在终端输出节点内容，并把节点信息写入到 /etc/dae/cnsub.sub。

./convert-sub-view.sh

查看 /etc/dae/cnsub.sub 内容

cat /etc/dae/cnsub.sub

节点信息类似下面的格式，

vless://auto%uuid@ip:端口?encryption=none&security=tls&sni=域名&fp=edge#🇳🇱 荷兰 re

在 dae 大鹅的配置 /etc/dae/config.dae 中，修改指定机场节点的名称关键词即可完成修改，例如。

    proxy {
        # 节点选择策略
        policy: min_moving_avg

        # 指定机场节点
        filter: subtag(my_sub) && name(keyword: '荷兰 re')
    }