起帆Cloud 深度评测：从购买套餐到客户端连接完整体验

mindmap
  root((起帆Cloud 2026 深度评测))
    品牌演进与协议特性
      2023年崛起平价代表
      Hysteria 2协议深度解析
      控制面板与购买流程
      现代客户端生态兼容
    套餐梯队与成本测算
      青铜2000GB超值体验
      白银6000GB边际单价
      黄金18000GB极限配额
      月付安全防坑准则
    网络拓扑与全球节点
      UDP智能拥塞控制
      全球核心出口机房
      节点倍率计费规则
      动态端口跳跃自愈
    晚高峰全天候测速
      千兆对称宽带测试
      晚高峰吞吐走势分析
      三网运营商延迟对比
      极端高压丢包与抖动
    流媒体与AI工具兼容
      Netflix与Disney解锁
      ChatGPT免验登录
      Cursor极低时延联想
      流式长连接不断流
    客户端调优与自检
      Clash Verge与Surge
      Singbox现代内核
      自动化监控脚本库
    真实案例与长周期评估
      突发超时故障切换
      720小时稳定性报告
      竞品横向对比与总结

timeline
    title 起帆Cloud 技术与品牌演变
    2023 : 平台正式创立 : 率先试水 Hysteria 协议与海量流量模式
    2024 : 升级 Hysteria 2 协议 : 降低握手开销并提升抗封锁能力
    2025 : 扩充全球机房出口 : 完善流媒体原生解锁与 AI 适配
    2026 : 维持高性价比运转 : 成为大流量下载与影视发烧友热门选型

graph TD
    SubPricing[起帆Cloud 套餐体系] --> BronzePlan[青铜 Bronze 套餐<br>月费 9.9 元 / 独享 2000GB 流量]
    SubPricing --> SilverPlan[白银 Silver 套餐<br>月费 24 元 / 独享 6000GB 流量]
    SubPricing --> GoldPlan[黄金 Gold 套餐<br>月费 60 元 / 独享 18000GB 流量]

flowchart TD
    UserTerminal[本地用户设备] --> LocalKernel[支持 Hy2 代理内核]
    LocalKernel -->|定制 UDP 拥塞控制 / 端口跳跃| PublicISP[公网骨干网络]
    PublicISP --> TransitEdge[直连落地或中继服务器集群]
    TransitEdge --> HK_Exit[香港优化节点]
    TransitEdge --> JP_Exit[日本高速机房]
    TransitEdge --> SG_Exit[新加坡亚太出口]
    TransitEdge --> US_Exit[美国大带宽机房]

graph LR
    ActualTraffic[用户实际拉取 100GB 4K 蓝光电影]
    ActualTraffic -->|连接 1.0 倍率标准出口| Billed[账单扣减 100GB]
    Billed --> Remaining[剩余 1900GB 配额充沛无忧]

graph TD
    TestBench[测试基准平台] --> SpecNet[千兆对称光纤 FTTH 直连万兆网卡]
    TestBench --> SpecTime[晚间 20:00 - 23:00 骨干拥塞黄金期]
    TestBench --> SpecTool[50 线程 UDP 暴力满载吞吐采集]
    TestBench --> SpecISP[电信 / 联通 / 移动三网真实物理接入]

xychart-beta
    title "晚高峰 20:00 - 23:00 主力节点下行速率表现 (Mbps)"
    x-axis ["20:00", "20:30", "21:00", "21:30", "22:00", "22:30", "23:00"]
    y-axis "下行带宽 (Mbps)" 0 --> 700
    line "美国 01 节点 (Hy2)" [620, 560, 480, 430, 470, 540, 610]
    line "日本 01 节点 (Hy2)" [450, 410, 360, 320, 350, 420, 460]
    line "香港 01 节点 (Hy2)" [430, 390, 340, 310, 330, 400, 440]

graph TD
    StreamQuery[终端发起 4K / 8K 视听请求] --> NodeLanding[起帆Cloud 出口集群]
    NodeLanding --> MediaInspect{流媒体平台版权校验}
    MediaInspect -->|通过原生 IP 认证| DirectPlay[秒开 4K 杜比全景声 码率顶格]
    MediaInspect -->|识别为普通云机房| LimitedAccess[仅限自制剧集 / 提示代理警告]

flowchart LR
    DevClient[开发者发起 API / 网页交互] --> QifanExit[起帆Cloud 指定出站节点]
    QifanExit --> WAF_Gate[OpenAI / Claude 安全网关]
    WAF_Gate -->|通过风控校验| ModelOutput[流式推演高速响应]
    WAF_Gate -->|触发 IP 拦截| AccessDenied[返回 403 访问受限]

graph TD
    DataFlow[客户端发起网络请求] --> CoreEngine{规则分流路由引擎}
    CoreEngine -->|国内域名与直连网段| DirectTraffic[DIRECT 本地直接连接]
    CoreEngine -->|OpenAI / Claude / Copilot| AIPolicy[专线 AI 专用路由组]
    CoreEngine -->|Netflix / Disney+ / YouTube| MediaPolicy[Hysteria2 流媒体大带宽组]
    CoreEngine -->|GitHub / 技术开发资源| CodePolicy[亚太低时延中转组]
    CoreEngine -->|未命中其余海外流量| FallbackGroup[自动故障切换主策略]

# Clash Verge Rev (Mihomo 内核) 生产级分流配置片段
mixed-port: 7890
allow-lan: false
mode: rule
log-level: info
ipv6: false

dns:
  enable: true
  listen: 127.0.0.1:1053
  ipv6: false
  enhanced-mode: fake-ip
  fake-ip-range: 198.18.0.1/16
  default-nameserver:
    - 223.5.5.5
    - 119.29.29.29
  nameserver:
    - https://dns.alidns.com/dns-query
    - https://doh.pub/dns-query
  fallback:
    - https://1.1.1.1/dns-query
    - https://8.8.8.8/dns-query
  fallback-filter:
    geoip: true
    geoip-code: CN
    ipcidr:
      - 240.0.0.0/4

proxy-groups:
  - name: 起帆Cloud-主策略
    type: select
    proxies:
      - 智能故障转移
      - 香港核心组
      - 日本核心组
      - 新加坡核心组
      - 美国核心组
      - DIRECT

  - name: 智能故障转移
    type: fallback
    url: https://www.gstatic.com/generate_204
    interval: 180
    proxies:
      - 香港核心组
      - 日本核心组
      - 新加坡核心组

  - name: 人工智能
    type: select
    proxies:
      - 美国核心组
      - 日本核心组
      - 新加坡核心组

  - name: 海外流媒体
    type: select
    proxies:
      - 美国核心组
      - 香港核心组
      - 日本核心组
      - 新加坡核心组

  - name: 香港核心组
    type: select
    include-all: true
    filter: "港|HK|HongKong"

  - name: 日本核心组
    type: select
    include-all: true
    filter: "日|JP|Japan"

  - name: 新加坡核心组
    type: select
    include-all: true
    filter: "新|SG|Singapore"

  - name: 美国核心组
    type: select
    include-all: true
    filter: "美|US|United"

rules:
  - DOMAIN-SUFFIX,openai.com,人工智能
  - DOMAIN-SUFFIX,chatgpt.com,人工智能
  - DOMAIN-SUFFIX,anthropic.com,人工智能
  - DOMAIN-SUFFIX,claude.ai,人工智能
  - DOMAIN-SUFFIX,cursor.sh,人工智能
  - DOMAIN-SUFFIX,cursor.com,人工智能
  - DOMAIN-SUFFIX,githubcopilot.com,人工智能
  - DOMAIN-KEYWORD,netflix,海外流媒体
  - DOMAIN-SUFFIX,netflix.com,海外流媒体
  - DOMAIN-SUFFIX,nflxext.com,海外流媒体
  - DOMAIN-SUFFIX,disneyplus.com,海外流媒体
  - DOMAIN-SUFFIX,youtube.com,海外流媒体
  - DOMAIN-SUFFIX,googlevideo.com,海外流媒体
  - GEOSITE,cn,DIRECT
  - GEOIP,CN,DIRECT
  - MATCH,起帆Cloud-主策略

[General]
loglevel = notify
skip-proxy = 127.0.0.1, 192.168.0.0/16, 10.0.0.0/8, 172.16.0.0/12, localhost, *.local, captive.apple.com
bypass-tun = 192.168.0.0/16, 10.0.0.0/8, 172.16.0.0/12
dns-server = 223.5.5.5, 119.29.29.29
encrypted-dns-server = https://dns.alidns.com/dns-query

[Proxy Group]
QifanCloud = select, Auto-Fallback, HK-Nodes, JP-Nodes, SG-Nodes, US-Nodes, DIRECT
Auto-Fallback = fallback, policy-path = https://sub.qifan.example/surge, url = http://cp.cloudflare.com/generate_204, interval = 300, timeout = 3
AI-Outbound = select, US-Nodes, JP-Nodes, SG-Nodes
Media-Streaming = select, US-Nodes, HK-Nodes, SG-Nodes, JP-Nodes
HK-Nodes = policy-path = https://sub.qifan.example/surge, policy-regex-filter = 港|HK
JP-Nodes = policy-path = https://sub.qifan.example/surge, policy-regex-filter = 日|JP
SG-Nodes = policy-path = https://sub.qifan.example/surge, policy-regex-filter = 新|SG
US-Nodes = policy-path = https://sub.qifan.example/surge, policy-regex-filter = 美|US

[Rule]
DOMAIN-SUFFIX,openai.com,AI-Outbound
DOMAIN-SUFFIX,chatgpt.com,AI-Outbound
DOMAIN-SUFFIX,claude.ai,AI-Outbound
DOMAIN-SUFFIX,anthropic.com,AI-Outbound
DOMAIN-SUFFIX,cursor.sh,AI-Outbound
DOMAIN-KEYWORD,netflix,Media-Streaming
DOMAIN-SUFFIX,netflix.com,Media-Streaming
DOMAIN-SUFFIX,disneyplus.com,Media-Streaming
DOMAIN-SUFFIX,youtube.com,Media-Streaming
RULE-SET,SYSTEM,DIRECT
GEOIP,CN,DIRECT
FINAL,QifanCloud,dns-failed

{
  "log": {
    "level": "info",
    "timestamp": true
  },
  "dns": {
    "servers": [
      {
        "tag": "dns-remote",
        "address": "https://1.1.1.1/dns-query",
        "detour": "proxy"
      },
      {
        "tag": "dns-direct",
        "address": "223.5.5.5",
        "detour": "direct"
      },
      {
        "tag": "dns-fake",
        "address": "fakeip"
      }
    ],
    "rules": [
      {
        "outbound": "any",
        "server": "dns-direct"
      },
      {
        "rule_set": "geosite-cn",
        "server": "dns-direct"
      }
    ],
    "fakeip": {
      "enabled": true,
      "inet4_range": "198.18.0.0/15"
    }
  },
  "inbounds": [
    {
      "type": "tun",
      "tag": "tun-in",
      "inet4_address": "172.19.0.1/30",
      "auto_route": true,
      "strict_route": true,
      "sniff": true
    }
  ],
  "outbounds": [
    {
      "type": "selector",
      "tag": "proxy",
      "outbounds": ["auto-fallback", "hk-hy2", "jp-hy2", "us-hy2"]
    },
    {
      "type": "urltest",
      "tag": "auto-fallback",
      "outbounds": ["hk-hy2", "jp-hy2"],
      "url": "https://www.gstatic.com/generate_204",
      "interval": "3m"
    },
    {
      "type": "selector",
      "tag": "ai-route",
      "outbounds": ["us-hy2", "jp-hy2"]
    },
    {
      "type": "direct",
      "tag": "direct"
    },
    {
      "type": "dns",
      "tag": "dns-out"
    }
  ],
  "route": {
    "rules": [
      {
        "protocol": "dns",
        "outbound": "dns-out"
      },
      {
        "domain_suffix": ["openai.com", "anthropic.com", "claude.ai", "cursor.sh"],
        "outbound": "ai-route"
      },
      {
        "rule_set": ["geosite-cn", "geoip-cn"],
        "outbound": "direct"
      }
    ],
    "auto_detect_interface": true
  }
}

flowchart LR
    SchedulerCron[定时计划任务] --> WorkerPing[TCP / UDP 握手时延测量]
    SchedulerCron --> WorkerLoss[连续丢包离散度统计]
    SchedulerCron --> WorkerUnlock[关键海外服务状态探针]
    WorkerPing --> LocalStorage[(本地指标日志数据库)]
    WorkerLoss --> LocalStorage
    WorkerUnlock --> LocalStorage
    LocalStorage --> GenerateReport[输出综合质量分析简报]

#!/usr/bin/env python3
import asyncio
import socket
import time
from typing import List, Tuple

TARGET_ENTRIES: List[Tuple[str, int, str]] = [
    ("hk01.qifan-node.net", 443, "香港 01 节点"),
    ("jp01.qifan-node.net", 443, "日本 01 节点"),
    ("sg01.qifan-node.net", 443, "新加坡 01 节点"),
    ("us01.qifan-node.net", 443, "美国 01 节点"),
]

async def probe_node(host: str, port: int, label: str) -> Tuple[str, float, bool]:
    start = time.perf_counter()
    try:
        reader, writer = await asyncio.wait_for(
            asyncio.open_connection(host, port), timeout=3.0
        )
        elapsed = (time.perf_counter() - start) * 1000.0
        writer.close()
        await writer.wait_closed()
        return label, elapsed, True
    except Exception:
        return label, 9999.0, False

async def main():
    print("开始执行起帆Cloud 主力入口节点批量握手时延测速...")
    tasks = [probe_node(h, p, l) for h, p, l in TARGET_ENTRIES]
    records = await asyncio.gather(*tasks)
    
    print("=" * 55)
    for label, rtt, status in records:
        stat_label = "正常" if status else "超时"
        print(f"[{stat_label}] {label:<16} 握手时延 {rtt:6.2f} ms")
    print("=" * 55)

if __name__ == "__main__":
    asyncio.run(main())

#!/usr/bin/env bash
# 起帆Cloud 晚高峰网络质量采样工具
TARGET_HOST="hk01.qifan-node.net"
TARGET_PORT="443"
TOTAL_SAMPLES=30
CONN_TIMEOUT=2

echo "正在对目标入口 ${TARGET_HOST}:${TARGET_PORT} 进行连续 ${TOTAL_SAMPLES} 次连接质量采样..."

SUCC=0
FAIL=0
SUM_TIME=0
DELAYS=()

for i in $(seq 1 $TOTAL_SAMPLES); do
    T_START=$(date +%s%N)
    if nc -z -w $CONN_TIMEOUT "$TARGET_HOST" "$TARGET_PORT" 2>/dev/null; then
        T_END=$(date +%s%N)
        DIFF=$(( (T_END - T_START) / 1000000 ))
        DELAYS+=("$DIFF")
        SUM_TIME=$(( SUM_TIME + DIFF ))
        SUCC=$(( SUCC + 1 ))
    else
        FAIL=$(( FAIL + 1 ))
    fi
    sleep 0.1
done

LOSS_PERCENT=$(awk "BEGIN {print ($FAIL / $TOTAL_SAMPLES) * 100}")
if [ $SUCC -gt 0 ]; then
    AVG_RTT=$(awk "BEGIN {print $SUM_TIME / $SUCC}")
else
    AVG_RTT=0
fi

echo "=================== 采样统计归档 ==================="
echo "总测试样本数: $TOTAL_SAMPLES"
echo "成功连接数:   $SUCC"
echo "失败超时数:   $FAIL"
echo "丢包率统计:   ${LOSS_PERCENT}%"
echo "平均握手时延: ${AVG_RTT} ms"
echo "==================================================="

#!/usr/bin/env python3
import requests
import sys

LOCAL_PROXY = {
    "http": "http://127.0.0.1:7890",
    "https": "http://127.0.0.1:7890"
}

AUDIT_TARGETS = [
    {
        "service": "OpenAI 鉴权接口",
        "url": "https://chatgpt.com/api/auth/session",
        "pass_codes": [200, 401]
    },
    {
        "service": "Claude 门户主页",
        "url": "https://claude.ai/login",
        "pass_codes": [200, 302]
    },
    {
        "service": "Netflix 视频检索",
        "url": "https://www.netflix.com/title/80018499",
        "pass_codes": [200]
    }
]

def check_services():
    print("正在通过本地代理测试关键服务连通性与风控状态...")
    session = requests.Session()
    session.proxies = LOCAL_PROXY
    session.headers.update({
        "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36"
    })

    for item in AUDIT_TARGETS:
        try:
            res = session.get(item["url"], timeout=6)
            if res.status_code in item["pass_codes"]:
                print(f"[通过] {item['service']:<16} 状态码 {res.status_code} 服务就绪")
            else:
                print(f"[受限] {item['service']:<16} 状态码 {res.status_code} 疑似受到拦截")
        except Exception as ex:
            print(f"[报错] {item['service']:<16} 请求超时 错误信息 {str(ex)[:35]}")

if __name__ == "__main__":
    check_services()

graph TD
    AlertOccur[突发连接故障或访问受阻] --> DiagnosticFlow{故障分类定位}
    DiagnosticFlow -->|本地运营商严格压制 UDP| SolutionCase1[启用备用端口跳跃与降级 TCP 节点]
    DiagnosticFlow -->|流媒体区域降级为自制剧| SolutionCase2[清除本地播放缓存并切换原生节点]
    DiagnosticFlow -->|终端命令行操作代理失效| SolutionCase3[配置终端全局代理环境变量]
    SolutionCase1 --> Restored[服务恢复正常]
    SolutionCase2 --> Restored
    SolutionCase3 --> Restored

pie title 720 小时全天候可用性追踪统计
    "高速平稳运行 99.15%" : 99.15
    "晚高峰轻微抖动 0.62%" : 0.62
    "机房割接维护 0.23%" : 0.23

radar
    title 起帆Cloud 与竞品核心能力对比
    "晚高峰吞吐" : 9.5
    "时延稳定性" : 8.2
    "流媒体解锁" : 8.6
    "节点丰富度" : 8.3
    "套餐性价比" : 9.8
    "运维持续度" : 8.5

graph LR
    UserNeed{核心使用场景} --> MovieLover[4K 蓝光影视发烧友]
    UserNeed --> Downloader[大文件与开源数据下载]
    UserNeed --> Developer[软件研发与技术办公]
    UserNeed --> EcomSeller[跨境电商与外贸运营]

    MovieLover --> Choice1[强烈推荐青铜 9.9 元套餐 极速秒开]
    Downloader --> Choice2[推荐白银或黄金套餐 享受海量流量]
    Developer --> Choice3[作为主力大吞吐通道 建议常备专线备用]
    EcomSeller --> Choice4[建议选用专属独立纯净 IP 保持谨慎]