<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"
  xmlns:atom="http://www.w3.org/2005/Atom"
  xmlns:dc="http://purl.org/dc/elements/1.1/"
  xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>Treeru 블로그</title>
    <link>https://treeru.com</link>
    <description>웹 개발, IT 인프라, AI 솔루션에 대한 인사이트와 실무 가이드</description>
    <language>ko</language>
    <managingEditor>admin@treeru.com (Treeru)</managingEditor>
    <webMaster>admin@treeru.com (Treeru)</webMaster>
    <lastBuildDate>Sat, 09 May 2026 15:33:26 GMT</lastBuildDate>
    <generator>Treeru Next.js RSS Generator</generator>
    <docs>https://www.rssboard.org/rss-specification</docs>
    <atom:link href="https://treeru.com/rss" rel="self" type="application/rss+xml"/>
    <image>
      <url>https://treeru.com/opengraph-image</url>
      <title>Treeru</title>
      <link>https://treeru.com</link>
      <width>144</width>
      <height>144</height>
    </image>
    <item>
      <title>웹 관리자 페이지 전체 스크린샷 자동화 — Playwright로 170장을 8분 만에</title>
      <link>https://treeru.com/blog/playwright-admin-ui-fullpage-screenshot-automation</link>
      <description>방화벽·라우터 같은 웹 관리자 UI의 모든 하위 페이지를 로그인 상태로 순회하며 fullPage PNG로 저장하는 방법을 정리합니다. Playwright 독립 스크립트로 170+ 페이지를 8분에 찍고, 자체서명 인증서 처리, 사이드바 순회, 날짜별 인덱스 생성까지 자동화합니다.</description>
      <pubDate>Tue, 21 Apr 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/playwright-admin-ui-fullpage-screenshot-automation</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>TOOL</category>
      <category>Playwright</category>
      <category>스크린샷</category>
      <category>자동화</category>
      <category>웹관리자UI</category>
      <category>백업</category>
      <category>CLI</category>
      <category>노드스크립트</category>
    </item>
    <item>
      <title>CQRS Documents by Greg Young (한글 번역 및 해설)</title>
      <link>https://treeru.com/blog/cqrs-documents-greg-young-korean</link>
      <description>Greg Young이 작성한 CQRS 백서 전체 56페이지를 한글로 완역한 블로그. CQRS, Event Sourcing, Task Based UI, Event Storage 구현까지 원문 문맥을 살려 페이지 단위로 번역했습니다. DDD와 이벤트 주도 아키텍처 학습용 한글 완역본입니다.</description>
      <pubDate>Tue, 14 Apr 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/cqrs-documents-greg-young-korean</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>CQRS</category>
      <category>DDD</category>
      <category>Event Sourcing</category>
      <category>Greg Young</category>
      <category>Domain Driven Design</category>
      <category>아키텍처</category>
      <category>번역</category>
    </item>
    <item>
      <title>EXAONE 4.5-33B RTX PRO 6000 으로 돌리면 어떨까 ?</title>
      <link>https://treeru.com/blog/exaone-45-33b-rtx-pro-6000-benchmark</link>
      <description>LG AI Research EXAONE 4.5-33B 출시 당일 RTX PRO 6000 Blackwell에서 벤치마크한 리포트. FP8 NaN crash, 커스텀 fork, BF16 22 TPS, 85문항 12시나리오 실전 결과와 안전성 100%까지 한국어 LLM 평가를 공개합니다.</description>
      <pubDate>Fri, 10 Apr 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/exaone-45-33b-rtx-pro-6000-benchmark</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>EXAONE</category>
      <category>LG AI Research</category>
      <category>RTX PRO 6000</category>
      <category>SGLang</category>
      <category>LLM벤치마크</category>
      <category>한국어AI</category>
      <category>VLM</category>
      <category>로컬AI</category>
      <category>Blackwell</category>
    </item>
    <item>
      <title>Zellij로 Claude Code SSH 환경 업그레이드 — tmux에서 갈아탄 이유</title>
      <link>https://treeru.com/blog/zellij-claude-code-ssh-session</link>
      <description>tmux 대신 Zellij로 Claude Code SSH 작업 환경을 구성한 실전 가이드. zj 자동화 스크립트, Windows bat 파일, 알려진 버그 사례, tmux 대비 리사이즈·복사·안정성 등 4가지 체감 차이까지 서버 여러 대를 원클릭으로 복귀하는 세션 관리법을 공개합니다.</description>
      <pubDate>Wed, 08 Apr 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/zellij-claude-code-ssh-session</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>CLAUDE-CODE</category>
      <category>zellij</category>
      <category>tmux</category>
      <category>Claude Code</category>
      <category>SSH</category>
      <category>터미널</category>
      <category>세션관리</category>
      <category>Linux</category>
      <category>서버관리</category>
      <category>자동화</category>
      <category>Windows Terminal</category>
    </item>
    <item>
      <title>OPNsense IDS를 AI로 30분 만에 실전 차단으로 — 방화벽 켜놓고 방치하지 않기</title>
      <link>https://treeru.com/blog/opnsense-ai-ids-security-automation</link>
      <description>OPNsense IDS를 켜놓고도 아무것도 안 막히는 이유 3가지와, AI와 REST API를 활용해 30분 만에 2,310개 DROP 규칙을 적용하고 실제 공격 22건을 차단한 과정을 공유합니다. Suricata 룰 최적화와 방화벽 보안 자동화를 다루는 실전 가이드입니다.</description>
      <pubDate>Mon, 06 Apr 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/opnsense-ai-ids-security-automation</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>네트워크</category>
      <category>OPNsense</category>
      <category>IDS</category>
      <category>방화벽</category>
      <category>보안자동화</category>
      <category>AI</category>
      <category>네트워크보안</category>
      <category>Suricata</category>
    </item>
    <item>
      <title>Gemma 4 출시 직후 RTX PRO 6000 실전 벤치마크 — 8번의 삽질과 Qwen3 비교</title>
      <link>https://treeru.com/blog/gemma4-rtx-pro-6000-benchmark</link>
      <description>Gemma 4 출시 2일 만에 RTX PRO 6000 Blackwell(96GB)에서 실전 벤치마크. SGLang과 vLLM 8번 삽질 끝에 AWQ로 성공. Gemma4-26B MoE가 Qwen3-32B보다 37% 빠른 결과와 3개 모델 속도·정확도 비교를 공개합니다.</description>
      <pubDate>Sun, 05 Apr 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/gemma4-rtx-pro-6000-benchmark</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>Gemma4</category>
      <category>RTX PRO 6000</category>
      <category>vLLM</category>
      <category>SGLang</category>
      <category>AWQ</category>
      <category>LLM벤치마크</category>
      <category>Qwen3</category>
      <category>로컬AI</category>
    </item>
    <item>
      <title>바이너리 로깅 입문 — 6개 포맷 비교와 선택 기준</title>
      <link>https://treeru.com/blog/avro-binary-logging-why-not-json</link>
      <description>바이너리 로깅 포맷 6종(MessagePack, Protobuf, CBOR, FlatBuffers, Avro, Cap&apos;n Proto)의 특성, 직렬화 속도, 파일 크기, 스키마 진화, AI 학습 데이터 적합성, 비동기 큐 통합까지 포맷별 장단점과 선택 기준을 종합 비교합니다.</description>
      <pubDate>Thu, 02 Apr 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/avro-binary-logging-why-not-json</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>로그</category>
      <category>Avro</category>
      <category>바이너리 로깅</category>
      <category>로그</category>
      <category>MessagePack</category>
      <category>Protobuf</category>
      <category>AI 학습</category>
      <category>스키마 진화</category>
      <category>비동기 큐</category>
    </item>
    <item>
      <title>15대 서버 임베딩 인프라 설계 — CPU 선택부터 아키텍처 분류까지</title>
      <link>https://treeru.com/blog/embedding-infrastructure-15-servers-cpu-selection</link>
      <description>15대 서버에 임베딩 인프라를 배치한 실전 사례. 9950X3D부터 N100까지 7종 CPU 성능 비교, RAM·NVMe 브랜드 무관 실증, LLM 파이프라인 내 임베딩 비중 6% 미만 분석, 서버 Tier 분류 체계까지 RAG 인프라 설계의 모든 것을 상세히 공개합니다.</description>
      <pubDate>Sun, 29 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/embedding-infrastructure-15-servers-cpu-selection</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>TOOL</category>
      <category>임베딩</category>
      <category>인프라설계</category>
      <category>CPU</category>
      <category>RAG</category>
      <category>AI서버</category>
      <category>서버아키텍처</category>
      <category>성능최적화</category>
    </item>
    <item>
      <title>Tmux + Claude Code — SSH 끊겨도 AI가 계속 일하는 환경 만들기</title>
      <link>https://treeru.com/blog/tmux-claude-code-ssh-persistent-session</link>
      <description>SSH 끊김으로 Claude Code 세션이 날아가는 문제를 tmux로 해결한 실전 가이드. 자동화 스크립트, .tmux.conf 설정, Windows bat 파일, 15대 서버 배포 설정, 트러블슈팅까지 네트워크 끊김에도 AI가 계속 일하는 환경 구축법을 공개합니다.</description>
      <pubDate>Sun, 29 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/tmux-claude-code-ssh-persistent-session</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>CLAUDE-CODE</category>
      <category>tmux</category>
      <category>Claude Code</category>
      <category>SSH</category>
      <category>터미널</category>
      <category>세션관리</category>
      <category>Linux</category>
      <category>서버관리</category>
      <category>자동화</category>
    </item>
    <item>
      <title>Microsoft Store 없이 winget + Windows Terminal 복구하고 Claude Code 설치하기</title>
      <link>https://treeru.com/blog/windows-terminal-winget-claude-code-install</link>
      <description>Microsoft Store가 고장난 Windows 11에서 VCLibs, UI.Xaml, App Installer를 수동으로 설치해 winget을 복구하고, Windows Terminal과 Claude Code까지 설치한 실전 가이드. 의존성 패키지와 설치 순서를 단계별로 공개합니다.</description>
      <pubDate>Sun, 29 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/windows-terminal-winget-claude-code-install</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>CLAUDE-CODE</category>
      <category>winget</category>
      <category>Windows Terminal</category>
      <category>Claude Code</category>
      <category>Windows 11</category>
      <category>PowerShell</category>
      <category>App Installer</category>
      <category>수동설치</category>
      <category>Microsoft Store</category>
    </item>
    <item>
      <title>AI 이미지 생성 API 2026년 가격 비교 — 블로그 썸네일 자동화 직접 써보고 정리</title>
      <link>https://treeru.com/blog/ai-image-generation-api-price-comparison-2026</link>
      <description>블로그 썸네일 자동화에 AI 이미지 API 1년 쓴 가격 비교 리포트. Gemini 2.5 Flash, GPT Image 1.5, FLUX 1.1 Pro, DALL-E 3 비용과 NO TEXT 신뢰도, OCR 재생성 안전망까지 장당 $0.009~$0.045 실사용 후기로 분석합니다.</description>
      <pubDate>Wed, 25 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/ai-image-generation-api-price-comparison-2026</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>AI이미지생성</category>
      <category>Gemini</category>
      <category>GPT Image</category>
      <category>FLUX</category>
      <category>이미지API</category>
      <category>블로그자동화</category>
      <category>가격비교</category>
    </item>
    <item>
      <title>Claude API로 블로그 자동 글쓰기 1년 — 실제 비용과 자연스러운 문체 만들기</title>
      <link>https://treeru.com/blog/claude-api-blog-automation-cost-1year</link>
      <description>Claude Sonnet 4로 블로그 글을 24/7 자동 생성한 지 1년. 1포스트당 약 80원, 월 90포스트에 약 7,200원까지 실제 API 비용 공개와 단어 반복 감지 로직, 구어체 프롬프트 기법, Sonnet 4.6 업그레이드 효과까지 운영 경험을 상세히 정리했습니다.</description>
      <pubDate>Wed, 25 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/claude-api-blog-automation-cost-1year</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>Claude API</category>
      <category>블로그자동화</category>
      <category>Anthropic</category>
      <category>AI글쓰기</category>
      <category>콘텐츠자동화</category>
      <category>파이썬</category>
    </item>
    <item>
      <title>Claude Code를 쓸 때 IDE 대신 터미널 하나로 충분한 이유 — TreeRU 실사용 워크플로우</title>
      <link>https://treeru.com/blog/claude-code-treeru-daily-workflow</link>
      <description>Claude Code CLI와 TreeRU 터미널 파일 탐색기를 조합한 실전 워크플로우를 소개합니다. 스크린샷 자동 저장, 사용량 모니터링, SSH 원격 개발, 멀티 프로젝트 전환까지 IDE 없이 터미널 하나로 개발·디버깅·배포를 해결하는 일상 개발 환경 구축법입니다.</description>
      <pubDate>Wed, 25 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/claude-code-treeru-daily-workflow</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>CLAUDE-CODE</category>
      <category>Claude Code</category>
      <category>TreeRU</category>
      <category>터미널</category>
      <category>워크플로우</category>
      <category>IDE대안</category>
      <category>CLI개발</category>
      <category>스크린샷</category>
      <category>SSH</category>
    </item>
    <item>
      <title>NVIDIA 드라이버 595 업그레이드 실측 - RTX PRO 6000 Blackwell AI 성능 변화</title>
      <link>https://treeru.com/blog/nvidia-driver-595-blackwell-upgrade</link>
      <description>590에서 595로 드라이버 업그레이드 후 RTX PRO 6000 Blackwell(96GB)의 AI 추론 성능 변화를 실측합니다. SGLang 토큰 속도, CUDA 연산 성능, 메모리 대역폭 비교, P-State 변화, Power Limit 초기화 주의사항까지 운영 팁을 정리합니다.</description>
      <pubDate>Wed, 25 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/nvidia-driver-595-blackwell-upgrade</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>NVIDIA</category>
      <category>드라이버업그레이드</category>
      <category>RTX PRO 6000</category>
      <category>Blackwell</category>
      <category>CUDA</category>
      <category>SGLang</category>
      <category>벤치마크</category>
      <category>로컬AI</category>
    </item>
    <item>
      <title>파이썬 프로그램 시작 78초를 20초로 — 로그 타임라인으로 병목 찾기</title>
      <link>https://treeru.com/blog/python-startup-speed-optimization-78s-to-20s</link>
      <description>N100 미니PC에서 파이썬 프로그램 시작에 78초 걸리던 것을 로그 타임라인 분석으로 병목을 찾아 20초로 단축한 실전 과정. 자동 로그인 대기 단축, GitHub API 호출 44회에서 5회로 축소, 불필요한 기능 비활성화까지 성능 최적화 방법을 상세히 공개합니다.</description>
      <pubDate>Wed, 25 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/python-startup-speed-optimization-78s-to-20s</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>파이썬</category>
      <category>성능최적화</category>
      <category>시작속도</category>
      <category>로그분석</category>
      <category>GitHub API</category>
      <category>N100</category>
    </item>
    <item>
      <title>셀레니움 네이버 캡차 2026년 해결법 — 삽질 7단계 후 찾은 $cdc_ 변수</title>
      <link>https://treeru.com/blog/selenium-naver-captcha-undetected-chromedriver</link>
      <description>1년간 잘 되던 셀레니움 네이버 자동 로그인이 갑자기 캡차가 뜨기 시작했습니다. UA 수정, stealth 스크립트, 영구 프로필, 랜덤 딜레이 등 7단계 삽질 끝에 찾은 진짜 원인은 $cdc_ 변수였습니다. undetected-chromedriver로 완전 해결한 실전 가이드입니다.</description>
      <pubDate>Wed, 25 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/selenium-naver-captcha-undetected-chromedriver</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>셀레니움</category>
      <category>네이버</category>
      <category>캡차</category>
      <category>undetected-chromedriver</category>
      <category>봇탐지</category>
      <category>파이썬</category>
      <category>자동화</category>
    </item>
    <item>
      <title>셀레니움 User-Agent 하드코딩하면 안 되는 이유 — Chrome 버전 불일치 봇 탐지</title>
      <link>https://treeru.com/blog/selenium-user-agent-hardcoding-bot-detection</link>
      <description>Chrome/114.0.0.0을 1년간 하드코딩해서 잘 쓰고 있었는데, Chrome 146 업데이트 후 갑자기 캡차가 터진 실제 사례. UA와 TLS 핑거프린트 버전 불일치가 봇 탐지를 유발하는 원리와 한 줄 수정으로 해결하는 방법을 실전 디버깅 과정과 함께 공개합니다.</description>
      <pubDate>Wed, 25 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/selenium-user-agent-hardcoding-bot-detection</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>셀레니움</category>
      <category>User-Agent</category>
      <category>봇탐지</category>
      <category>Chrome</category>
      <category>TLS핑거프린트</category>
      <category>파이썬</category>
      <category>자동화</category>
    </item>
    <item>
      <title>undetected-chromedriver를 subprocess로 넘기려다 하루 날린 이야기</title>
      <link>https://treeru.com/blog/undetected-chromedriver-subprocess-issue</link>
      <description>subprocess에서 UC로 Chrome을 만들어 메인 프로세스에 debug port로 넘기려 했다. --remote-debugging-port 무시, detach 미지원, stdout 파이프 상속 문제, 보이지 않는 탭 이슈까지 — 하루 동안 겪은 4가지 문제와 최종 결론.</description>
      <pubDate>Wed, 25 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/undetected-chromedriver-subprocess-issue</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>undetected-chromedriver</category>
      <category>subprocess</category>
      <category>셀레니움</category>
      <category>Chrome</category>
      <category>파이썬</category>
      <category>봇탐지</category>
    </item>
    <item>
      <title>TreeRU — 터미널에서 파일 탐색과 스크린샷 붙여넣기를 한 번에</title>
      <link>https://treeru.com/blog/treeru-terminal-file-explorer</link>
      <description>AI CLI 도구 사용 시 불편한 파일 탐색과 이미지 전달을 해결하는 터미널 파일 탐색기 TreeRU 소개. 클립보드 스크린샷 자동 저장, 단축키 경로 복사, SSH 원격 탐색, 설치 30초, 최소 리소스, 오픈소스로 공개된 Claude Code 최적 워크플로우 도구입니다.</description>
      <pubDate>Fri, 20 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/treeru-terminal-file-explorer</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>TOOL</category>
      <category>TreeRU</category>
      <category>터미널</category>
      <category>파일탐색기</category>
      <category>CLI</category>
      <category>스크린샷</category>
      <category>SSH</category>
      <category>오픈소스</category>
    </item>
    <item>
      <title>대규모 시계열 데이터를 위한 DB 엔진 8종 비교 — 2,200만 건 실측 벤치마크</title>
      <link>https://treeru.com/blog/timeseries-database-comparison-8-engines</link>
      <description>2,200만 건 시계열 데이터로 8개 DB 엔진을 벤치마크한 실측 결과. QuestDB 25ms, ClickHouse 547ms, TimescaleDB 1,021ms 등 집계 쿼리 속도, 수집 처리량, 압축률까지 6가지 평가 기준으로 종합 분석한 시계열 DB 선택 가이드입니다.</description>
      <pubDate>Tue, 10 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/timeseries-database-comparison-8-engines</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>스토리지</category>
      <category>시계열DB</category>
      <category>TimescaleDB</category>
      <category>ClickHouse</category>
      <category>QuestDB</category>
      <category>DuckDB</category>
      <category>데이터베이스비교</category>
      <category>OLAP</category>
      <category>벤치마크</category>
    </item>
    <item>
      <title>DB+RAG 하이브리드 검색 — 팩트 정확도를 5배 올리는 법</title>
      <link>https://treeru.com/blog/hybrid-search-db-rag-fact-accuracy-5x</link>
      <description>DB+RAG 병렬 검색으로 LLM 팩트 정확도를 17.5%에서 92.5%로 끌어올린 실전 아키텍처. 4-Case 결과 조립 로직, 병렬 검색 18ms 오버헤드, PostgreSQL+Qdrant 통합, 스키마 매칭 95%, AI 자동 구조화까지 실측 벤치마크와 함께 공개합니다.</description>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/hybrid-search-db-rag-fact-accuracy-5x</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>하이브리드검색</category>
      <category>DB검색</category>
      <category>RAG</category>
      <category>병렬처리</category>
      <category>팩트정확도</category>
      <category>AI챗봇</category>
      <category>Qdrant</category>
      <category>PostgreSQL</category>
    </item>
    <item>
      <title>리눅스 서버 신규 설치 체크리스트 — BIOS부터 SMART 알림까지 8단계</title>
      <link>https://treeru.com/blog/linux-server-setup-checklist-8-steps</link>
      <description>새 리눅스 서버를 설치할 때 빠뜨리기 쉬운 8가지 설정을 체크리스트로 정리합니다. BIOS 자동 부팅, SSH 키 인증, kernel.panic 설정, journald 영구 로그, smartmontools 알림까지 16대 서버 운영에서 검증된 Ubuntu 서버 설치 절차를 공개합니다.</description>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/linux-server-setup-checklist-8-steps</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>TOOL</category>
      <category>리눅스서버</category>
      <category>서버설정</category>
      <category>SSH키인증</category>
      <category>smartmontools</category>
      <category>kernel.panic</category>
      <category>journald</category>
      <category>체크리스트</category>
      <category>Ubuntu</category>
    </item>
    <item>
      <title>LLM 디코딩 파라미터 비교 — Temperature 0.1~0.9 실전 분석</title>
      <link>https://treeru.com/blog/llm-decoding-parameter-temperature-comparison</link>
      <description>60문항 × 5단계 Temperature = 300회 실험. 고객 서비스엔 0.1, 창작엔 0.7이 최적일까? 환각 트랩에서 Temperature별 행동 차이, 시나리오별 최적값, RTX PRO 6000 vs 5060 Ti GPU 비교까지 LLM 디코딩 파라미터 튜닝을 실측합니다.</description>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/llm-decoding-parameter-temperature-comparison</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>디코딩파라미터</category>
      <category>Temperature</category>
      <category>환각방어</category>
      <category>LLM튜닝</category>
      <category>Qwen3</category>
      <category>프롬프트엔지니어링</category>
      <category>SGLang</category>
      <category>로컬AI</category>
    </item>
    <item>
      <title>로컬 LLM 6종 종합 벤치마크 — 60문항 한국어 실무 점수표</title>
      <link>https://treeru.com/blog/local-llm-6-model-korean-benchmark</link>
      <description>Qwen3-14B, Gemma-12B, KORMo-10B, Qwen3-8B, Phi-4, Llama-3.1-8B 등 6개 로컬 LLM을 60문항 7개 시나리오로 종합 비교합니다. 환각 트랩 통과율, 치명적 결함, 모델 크기별 tok/s 대비 품질 트레이드오프까지 실측한 벤치마크입니다.</description>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/local-llm-6-model-korean-benchmark</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>로컬LLM</category>
      <category>한국어벤치마크</category>
      <category>모델비교</category>
      <category>Qwen3</category>
      <category>Gemma</category>
      <category>KORMo</category>
      <category>AI챗봇</category>
      <category>환각테스트</category>
    </item>
    <item>
      <title>로컬 RAG 파이프라인 구축기 — 임베딩 모델 선정부터 환각 제거까지</title>
      <link>https://treeru.com/blog/local-rag-pipeline-embedding-to-hallucination-fix</link>
      <description>BGE-M3 vs mE5-Large 한국어 검색 품질 비교(Recall 100% vs 86.7%), Qdrant Top-3 정확도 100%, RAG 적용 시 아메리카노 가격 1,500원→4,500원 환각 제거까지 실측한 로컬 RAG 파이프라인 구축의 각 단계를 검증한 실전 가이드입니다.</description>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/local-rag-pipeline-embedding-to-hallucination-fix</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>RAG</category>
      <category>벡터DB</category>
      <category>Qdrant</category>
      <category>BGE-M3</category>
      <category>임베딩</category>
      <category>환각제거</category>
      <category>로컬AI</category>
      <category>LLM</category>
    </item>
    <item>
      <title>MoE vs Dense 실전 비교 — Qwen3-30B-A3B가 14B보다 느린 이유</title>
      <link>https://treeru.com/blog/moe-vs-dense-qwen3-30b-14b-comparison</link>
      <description>Qwen3-30B-A3B(MoE)와 14B(Dense)를 직접 비교. VRAM 56.9GB vs 27.5GB, 속도 22 vs 41 tok/s 차이와 KTransformers 하이브리드 추론 실패 원인을 18,432개 가중치 조사로 밝힌 MoE vs Dense 실전 디버깅 리포트입니다.</description>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/moe-vs-dense-qwen3-30b-14b-comparison</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>MoE</category>
      <category>Dense모델</category>
      <category>Qwen3-30B</category>
      <category>KTransformers</category>
      <category>GGUF양자화</category>
      <category>하이브리드추론</category>
      <category>AI디버깅</category>
      <category>로컬LLM</category>
    </item>
    <item>
      <title>다중 서버 백업·복구 전략 — rsync Pull 기반 자동 백업과 재해 복구</title>
      <link>https://treeru.com/blog/multi-server-backup-rsync-pull-disaster-recovery</link>
      <description>rsync Pull 방식으로 다중 서버를 자동 백업하고, flock 동시 실행 방지, SMART 모니터링으로 디스크 건강을 감시하며, 재해 시 빈 서버에 재구축하는 실전 복구 절차를 공개합니다. 16대 서버 운영 경험을 바탕으로 한 백업·재해 복구의 완전한 가이드입니다.</description>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/multi-server-backup-rsync-pull-disaster-recovery</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>스토리지</category>
      <category>서버백업</category>
      <category>rsync</category>
      <category>재해복구</category>
      <category>SMART모니터링</category>
      <category>cron</category>
      <category>WDRedPro</category>
      <category>콜드백업</category>
      <category>NFS</category>
    </item>
    <item>
      <title>전서버 NVMe 스토리지 벤치마크 — Optane부터 HDD까지 20개 디바이스 비교</title>
      <link>https://treeru.com/blog/nvme-storage-benchmark-20-devices</link>
      <description>14개 NVMe SSD와 6개 HDD를 fio로 벤치마크한 실측 결과. Sequential Read 6,193~254MB/s, Optane 905P QD1 IOPS가 NAND 대비 3.8배 빠른 이유, AI 서버 스토리지 선택 기준을 20개 디바이스 비교 데이터로 공개합니다.</description>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/nvme-storage-benchmark-20-devices</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>하드웨어</category>
      <category>NVMe</category>
      <category>SSD벤치마크</category>
      <category>Optane</category>
      <category>fio</category>
      <category>AI스토리지</category>
      <category>서버하드웨어</category>
      <category>IOPS</category>
      <category>HDD비교</category>
    </item>
    <item>
      <title>Qwen3-32B-AWQ 60문항 한국어 테스트 — 14B 대비 2배 느린데 품질은?</title>
      <link>https://treeru.com/blog/qwen3-32b-awq-korean-test-vs-14b</link>
      <description>Qwen3-32B-AWQ와 14B-AWQ를 동일 60문항, 7개 시나리오로 비교한 실측 리뷰. 69 vs 135 tok/s 속도 2배 차이에서 품질은 평균 5.4% 향상에 그치고, 법률 추론만 유의미한 이점을 보였습니다. 환각 트랩 통과율과 think 토큰까지 상세 분석했습니다.</description>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/qwen3-32b-awq-korean-test-vs-14b</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>Qwen3-32B</category>
      <category>Qwen3-14B</category>
      <category>모델크기비교</category>
      <category>한국어테스트</category>
      <category>SGLang</category>
      <category>AWQ</category>
      <category>로컬LLM</category>
      <category>벤치마크</category>
    </item>
    <item>
      <title>Text2SQL 실전 테스트 — LLM이 SQL을 직접 작성하면</title>
      <link>https://treeru.com/blog/text2sql-llm-router-real-world-test</link>
      <description>LLM 라우터가 자연어를 분류하고, Text2SQL이 13테이블 PostgreSQL에서 SQL을 생성합니다. 40문항 4-Tier 난이도에서 LLM만 10.8%, 동의어사전 41.2%, Text2SQL 70.1% 정확도를 실측해 DB 자연어 질의 아키텍처를 검증한 실전 가이드입니다.</description>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/text2sql-llm-router-real-world-test</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>Text2SQL</category>
      <category>LLM라우터</category>
      <category>자연어쿼리</category>
      <category>데이터베이스</category>
      <category>AI챗봇</category>
      <category>PostgreSQL</category>
      <category>Qwen3</category>
      <category>SQL생성</category>
    </item>
    <item>
      <title>AWQ 양자화 속도 비교 — 모델 크기별 INT4 vs BF16, 그리고 MoE의 반전</title>
      <link>https://treeru.com/blog/awq-quantization-speed-comparison</link>
      <description>RTX PRO 6000에서 16개 모델의 tok/s 속도를 실측했습니다. AWQ INT4 양자화로 4B 모델 1.88배, 32B 모델 2.94배 속도 향상을 확인했고 MoE 모델 Qwen3-30B가 14B Dense보다 빠른 반전까지 벤치마크 데이터로 종합 분석합니다.</description>
      <pubDate>Thu, 26 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/awq-quantization-speed-comparison</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>AWQ</category>
      <category>양자화</category>
      <category>INT4</category>
      <category>BF16</category>
      <category>MoE</category>
      <category>벤치마크</category>
      <category>SGLang</category>
      <category>토큰속도</category>
      <category>LLM최적화</category>
    </item>
    <item>
      <title>Grafana + Prometheus로 16대 서버 통합 모니터링 — 설치부터 알림까지</title>
      <link>https://treeru.com/blog/grafana-prometheus-16-server-monitoring</link>
      <description>16대 서버를 Grafana와 Prometheus로 통합 모니터링한 실전 구축기. node_exporter 설치, GPU 모니터링, CPU·메모리·온도·디스크 메트릭, 알림 규칙 설정, 대시보드 설계, 장기 운영 팁까지 서버 통합 감시 시스템 구축의 전 과정을 공개합니다.</description>
      <pubDate>Thu, 26 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/grafana-prometheus-16-server-monitoring</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>TOOL</category>
      <category>Grafana</category>
      <category>Prometheus</category>
      <category>서버모니터링</category>
      <category>node_exporter</category>
      <category>GPU모니터링</category>
      <category>대시보드</category>
      <category>알림설정</category>
    </item>
    <item>
      <title>SGLang 23모델 서빙 실전 가이드 — 모델별 최적 설정 총정리</title>
      <link>https://treeru.com/blog/sglang-multi-model-serving-guide</link>
      <description>SGLang으로 23개 LLM을 실제 서빙하며 정리한 모델별 최적 설정. AWQ/BF16/FP8 양자화 선택 기준, MoE flashinfer 필수 설정, VRAM 관리, OOM 대응, 자동화 스크립트, 트러블슈팅까지 프로덕션 환경에서 검증한 SGLang 운영 실전 노하우를 공개합니다.</description>
      <pubDate>Thu, 26 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/sglang-multi-model-serving-guide</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>SGLang</category>
      <category>모델서빙</category>
      <category>AWQ</category>
      <category>BF16</category>
      <category>MoE</category>
      <category>VRAM</category>
      <category>LLM운영</category>
      <category>서빙최적화</category>
      <category>트러블슈팅</category>
    </item>
    <item>
      <title>Intel Optane 905P + NVMe 3-Tier 스토리지 — AI 서버 디스크 전략</title>
      <link>https://treeru.com/blog/ai-server-storage-strategy-optane</link>
      <description>AI 서버에 Intel Optane 905P(10μs 랜덤 읽기) + Samsung 980 PRO + NVMe로 3-Tier 스토리지를 구성한 실전 전략. RAG 벡터DB 배치, 모델 보관, NFS Pull 콜드백업까지 용도별 디스크 배치와 데이터 흐름, 실측 성능 차이를 공개합니다.</description>
      <pubDate>Tue, 24 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/ai-server-storage-strategy-optane</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>스토리지</category>
      <category>Optane</category>
      <category>스토리지</category>
      <category>3-Tier</category>
      <category>RAG</category>
      <category>NVMe</category>
      <category>콜드백업</category>
    </item>
    <item>
      <title>RTX 5060 Ti vs RTX PRO 6000 — 가격 11배 차이, 성능은?</title>
      <link>https://treeru.com/blog/rtx-5060-ti-vs-rtx-pro-6000-cost-performance</link>
      <description>RTX 5060 Ti($450)와 RTX PRO 6000($5,000)의 실 추론 속도, 동시접속 처리량, 안정성, 전력 효율을 직접 비교합니다. 가격의 9%로 성능의 35%를 추가하는 멀티 GPU 가성비 전략과 각 GPU의 실전 투입 기준을 실측 벤치마크 데이터로 분석합니다.</description>
      <pubDate>Mon, 23 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/rtx-5060-ti-vs-rtx-pro-6000-cost-performance</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>하드웨어</category>
      <category>RTX 5060 Ti</category>
      <category>RTX PRO 6000</category>
      <category>GPU비교</category>
      <category>가성비</category>
      <category>로컬AI</category>
      <category>멀티GPU</category>
    </item>
    <item>
      <title>RTX 5060 Ti 로컬 AI 종합 벤치마크 - $450 GPU의 실력</title>
      <link>https://treeru.com/blog/rtx-5060-ti-local-llm-benchmark</link>
      <description>RTX 5060 Ti(16GB)에서 8B·14B 모델의 단일 속도, 동시 30명 부하, 멀티턴 채팅, 크로스서버 추론까지 실측한 종합 벤치마크. 8B 모델 76 tok/s, 30명 동시 에러 0건 등 $450 GPU로 로컬 AI 서비스가 실제 가능한지 검증한 데이터 리포트입니다.</description>
      <pubDate>Sat, 21 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/rtx-5060-ti-local-llm-benchmark</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>하드웨어</category>
      <category>RTX 5060 Ti</category>
      <category>GPU벤치마크</category>
      <category>로컬AI</category>
      <category>SGLang</category>
      <category>동시접속</category>
      <category>가성비GPU</category>
    </item>
    <item>
      <title>Qwen3-14B 심층 리뷰 — 왜 1순위 추천 모델인가</title>
      <link>https://treeru.com/blog/qwen3-14b-deep-review</link>
      <description>Qwen3-14B-AWQ를 두 GPU에서 동일 60문항, 7가지 시나리오로 테스트한 심층 리뷰. 종합 3.86점 1위의 실력과 법조문 날조, 환각, 중국어 오염 등 한계까지 솔직하게 분석합니다. 1순위 추천 모델의 진짜 실력과 실무 투입 시 주의점을 모두 정리했습니다.</description>
      <pubDate>Thu, 19 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/qwen3-14b-deep-review</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>Qwen3-14B</category>
      <category>모델리뷰</category>
      <category>한국어AI</category>
      <category>환각테스트</category>
      <category>로컬AI</category>
      <category>LLM평가</category>
    </item>
    <item>
      <title>블로그 SEO 실전 — 내부 링크, relatedPosts, 시리즈 네비게이션</title>
      <link>https://treeru.com/blog/blog-seo-internal-linking-strategy</link>
      <description>블로그 글을 유기적으로 연결하는 내부 링크 전략을 공개합니다. relatedPosts 상호 연결, 시리즈 네비게이션, 카테고리 페이지, RSS feed 자동 생성, 앵커 텍스트 최적화까지 SEO와 체류시간을 동시에 개선하는 블로그 내부 링크 설계 및 구현 실전 가이드입니다.</description>
      <pubDate>Wed, 18 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/blog-seo-internal-linking-strategy</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>SEO</category>
      <category>내부링크</category>
      <category>SEO</category>
      <category>relatedPosts</category>
      <category>시리즈네비게이션</category>
      <category>블로그SEO</category>
      <category>RSS</category>
    </item>
    <item>
      <title>크로스서버 AI 추론 — 보조 GPU로 처리량 70% 올리기</title>
      <link>https://treeru.com/blog/cross-server-ai-inference-multi-gpu</link>
      <description>메인 서버(RTX PRO 6000)에 보조 서버(RTX 5060 Ti)를 1GbE로 연결해 AI 처리량을 70% 올린 실측 데이터를 공유합니다. 네트워크 오버헤드 5~18%, 라우팅 전략, 비용 대비 효율, 실전 운영 시 주의점까지 저비용 GPU 확장 실전 가이드입니다.</description>
      <pubDate>Wed, 18 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/cross-server-ai-inference-multi-gpu</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>크로스서버</category>
      <category>멀티GPU</category>
      <category>처리량</category>
      <category>라우팅</category>
      <category>로컬AI</category>
      <category>가성비</category>
      <category>분산추론</category>
    </item>
    <item>
      <title>RTX 5090 vs RTX PRO 6000 AI 추론 속도 비교 - 어떤 GPU가 로컬 LLM에 유리할까?</title>
      <link>https://treeru.com/blog/rtx-5090-vs-rtx-pro-6000-ai-benchmark</link>
      <description>RTX 5090(32GB)과 RTX PRO 6000(96GB)의 로컬 LLM 추론 속도를 llama-bench로 비교합니다. Qwen 32B·Llama 70B·Qwen 72B의 프롬프트·토큰 속도, GPU 스트레스 테스트, VRAM이 70B+ 모델 구동에 미치는 영향을 분석합니다.</description>
      <pubDate>Mon, 16 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/rtx-5090-vs-rtx-pro-6000-ai-benchmark</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>하드웨어</category>
      <category>RTX 5090</category>
      <category>RTX PRO 6000</category>
      <category>GPU벤치마크</category>
      <category>로컬AI</category>
      <category>LLM추론속도</category>
    </item>
    <item>
      <title>PageSpeed Insights API 자동화 — 전체 사이트 성능을 한 눈에</title>
      <link>https://treeru.com/blog/pagespeed-insights-api-automation</link>
      <description>매번 수동으로 PageSpeed 돌리지 마세요. PageSpeed Insights API v5를 활용한 전체 페이지 순차 측정, 마크다운 리포트 자동 생성, 관리자 UI 원클릭 측정, 장기 성능 트렌드 추적까지 사이트 성능을 한눈에 파악하는 자동화 도구 구축 가이드입니다.</description>
      <pubDate>Sun, 15 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/pagespeed-insights-api-automation</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>PageSpeed API</category>
      <category>자동화</category>
      <category>성능측정</category>
      <category>Lighthouse</category>
      <category>모니터링</category>
      <category>DevOps</category>
    </item>
    <item>
      <title>GPU 1장으로 7개 회사 동시 서빙 — 멀티테넌트 격리 실전 테스트</title>
      <link>https://treeru.com/blog/multi-tenant-ai-isolation-test</link>
      <description>LoRA 어댑터 5개로 7개 회사를 GPU 1장에서 동시 서빙하며 KV 캐시 누출, 어댑터 격리, 동시 페르소나, 프롬프트 인젝션 등 15건의 격리 테스트를 실행했습니다. 전부 PASS한 실측 결과와 SGLang 기반 멀티테넌트 아키텍처 설계 노하우를 상세히 공개합니다.</description>
      <pubDate>Sat, 14 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/multi-tenant-ai-isolation-test</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>멀티테넌트</category>
      <category>격리테스트</category>
      <category>LoRA</category>
      <category>KV캐시</category>
      <category>보안</category>
      <category>로컬AI</category>
      <category>SGLang</category>
    </item>
    <item>
      <title>Next.js SEO 100점 만들기 — 메타데이터, sitemap, robots.txt</title>
      <link>https://treeru.com/blog/nextjs-seo-100-metadata-sitemap-robots</link>
      <description>Next.js에서 SEO 100점을 만드는 가이드. generateMetadata, OG 이미지 자동 생성, 동적 sitemap.xml, robots.txt, JSON-LD 구조화 데이터, IndexNow, hreflang까지 Next.js 15 기준 SEO의 모든 것을 공개합니다.</description>
      <pubDate>Fri, 13 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/nextjs-seo-100-metadata-sitemap-robots</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>SEO</category>
      <category>Next.js</category>
      <category>SEO</category>
      <category>메타데이터</category>
      <category>sitemap</category>
      <category>robots.txt</category>
      <category>JSON-LD</category>
    </item>
    <item>
      <title>GPU 24/7 장기 운영 모니터링 — 13일간 실측 데이터 분석</title>
      <link>https://treeru.com/blog/gpu-long-term-monitoring-data</link>
      <description>RTX PRO 6000을 13일간 24/7 모니터링한 3,667개 데이터 포인트 분석 리포트. 유휴 11°C부터 부하 73°C, 600W→350W 전력 제한 전후 비교, VRAM 87.6% 상시 점유 패턴까지 AI 서버 장기 운영의 온도, 전력, 안정성을 실측 데이터로 검증합니다.</description>
      <pubDate>Thu, 12 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/gpu-long-term-monitoring-data</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>하드웨어</category>
      <category>GPU모니터링</category>
      <category>장기운영</category>
      <category>RTX PRO 6000</category>
      <category>온도관리</category>
      <category>전력제한</category>
      <category>안정성</category>
    </item>
    <item>
      <title>NFS 콜드백업 서버 구축 — IronWolf 12TB로 데이터 보호</title>
      <link>https://treeru.com/blog/nfs-cold-backup-server-ironwolf</link>
      <description>저전력 CPU와 IronWolf 12TB HDD로 NFS 콜드백업 서버를 구축하는 실전 과정. RAID 1 미러링, rsync와 cron 자동 백업, 듀얼 NIC로 백업 트래픽 격리, SMART 모니터링, 장애 대응까지 비용 효율적인 데이터 보호 시스템 구축 가이드입니다.</description>
      <pubDate>Wed, 11 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/nfs-cold-backup-server-ironwolf</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>스토리지</category>
      <category>NFS</category>
      <category>백업</category>
      <category>IronWolf</category>
      <category>데이터보호</category>
      <category>서버운영</category>
    </item>
    <item>
      <title>CPU 임베딩 벤치마크 도구 — 8종 CPU 성능 비교 자동화</title>
      <link>https://treeru.com/blog/cpu-embedding-benchmark-tool</link>
      <description>CPU마다 임베딩 속도가 이렇게 다릅니다. Python 벤치마크 스크립트로 9950X3D, 7840HS, N100 등 8종 CPU의 임베딩 성능을 자동 측정합니다. 최대 6.85배 속도 차이, SSH 원격 실행, JSON 출력, 와트당 효율 분석까지 AI 인프라 선택 가이드입니다.</description>
      <pubDate>Tue, 10 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/cpu-embedding-benchmark-tool</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>TOOL</category>
      <category>CPU벤치마크</category>
      <category>임베딩</category>
      <category>Python</category>
      <category>성능비교</category>
      <category>AI</category>
      <category>자동화</category>
    </item>
    <item>
      <title>GPU 전력 제한과 AI 성능 비교 - 언더볼팅·와트 제한 실측 데이터</title>
      <link>https://treeru.com/blog/gpu-power-limit-ai-performance</link>
      <description>RTX 5090 언더볼팅과 RTX PRO 6000 전력 제한(600W→450W→350W)이 AI 추론 속도와 GPU 온도에 미치는 영향을 실측합니다. 토큰 생성 속도 -3.8%, 온도 -22°C, 동시 부하 테스트 데이터로 전기료 절감과 성능의 최적 균형점을 찾는 분석 리포트입니다.</description>
      <pubDate>Tue, 10 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/gpu-power-limit-ai-performance</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>하드웨어</category>
      <category>GPU전력제한</category>
      <category>언더볼팅</category>
      <category>RTX PRO 6000</category>
      <category>로컬AI</category>
      <category>전기료절감</category>
    </item>
    <item>
      <title>Core Web Vitals 실전 디버깅 — Chrome DevTools로 CLS, LCP 원인 찾기</title>
      <link>https://treeru.com/blog/core-web-vitals-debugging-chrome-devtools</link>
      <description>PageSpeed 점수만 보지 말고 원인을 추적하세요. Chrome DevTools Performance 탭, Layout Shift 하이라이트, LCP 요소 식별, curl SSR 분석까지 CLS와 LCP 문제의 진짜 원인을 찾는 Core Web Vitals 디버깅 실전 가이드입니다.</description>
      <pubDate>Sun, 08 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/core-web-vitals-debugging-chrome-devtools</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>Core Web Vitals</category>
      <category>Chrome DevTools</category>
      <category>CLS</category>
      <category>LCP</category>
      <category>디버깅</category>
      <category>PageSpeed</category>
    </item>
    <item>
      <title>LoRA 파인튜닝으로 회사별 AI 만들기 — 학습부터 서빙까지</title>
      <link>https://treeru.com/blog/lora-finetuning-custom-ai-chatbot</link>
      <description>10쌍의 학습 데이터와 6초의 학습만으로 5개 회사별 AI 챗봇을 만든 실전 가이드. LoRA 어댑터 학습, 73MB 경량 어댑터, SGLang 핫스왑 서빙, 시스템 프롬프트 설계, 멀티테넌트 아키텍처까지 GPU 1장으로 다기업 커스텀 AI를 운영하는 전 과정을 공개합니다.</description>
      <pubDate>Sat, 07 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/lora-finetuning-custom-ai-chatbot</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>LoRA</category>
      <category>파인튜닝</category>
      <category>멀티테넌트</category>
      <category>챗봇</category>
      <category>SGLang</category>
      <category>로컬AI</category>
      <category>커스텀AI</category>
    </item>
    <item>
      <title>Next.js로 관리자 페이지 만들기 — 실전 어드민 패널 구축기</title>
      <link>https://treeru.com/blog/nextjs-admin-panel-development</link>
      <description>자체 서비스를 운영한다면 관리자 페이지는 필수입니다. Next.js 어드민 패널 아키텍처, 사이드바 레이아웃, SEO 관리, PageSpeed 통합, 인증 미들웨어, RESTful API 라우트 설계, 외부 서비스 연동까지 실전에서 쓰는 프로덕션급 어드민 구축 가이드입니다.</description>
      <pubDate>Fri, 06 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/nextjs-admin-panel-development</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>Next.js</category>
      <category>관리자페이지</category>
      <category>어드민</category>
      <category>API</category>
      <category>미들웨어</category>
      <category>웹개발</category>
    </item>
    <item>
      <title>SGLang vs vLLM 실전 비교 — 처리량 3배 차이의 비밀</title>
      <link>https://treeru.com/blog/sglang-vs-vllm-serving-engine-comparison</link>
      <description>Qwen3-32B와 LoRA 어댑터 5개 동일 조건에서 SGLang과 vLLM을 200명까지 부하 테스트한 실전 비교. 처리량 3배, P95 지연 6.5배 차이의 원인 분석과 에러율 0% 달성 조건, 실전 선택 기준을 실측 데이터와 함께 정리한 서빙 엔진 가이드입니다.</description>
      <pubDate>Thu, 05 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/sglang-vs-vllm-serving-engine-comparison</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>SGLang</category>
      <category>vLLM</category>
      <category>서빙엔진</category>
      <category>처리량</category>
      <category>P95</category>
      <category>로컬AI</category>
      <category>벤치마크</category>
    </item>
    <item>
      <title>PageSpeed 점수의 진실 — 측정할 때 알아야 할 것들</title>
      <link>https://treeru.com/blog/pagespeed-score-measurement-guide</link>
      <description>PageSpeed 점수는 같은 사이트도 매번 다릅니다. ±3~5점 변동 원인, Lighthouse 가중치, Mobile과 Desktop 차이, 측정 환경 변수, 90점 이후 투자 대비 효과가 급감하는 구간까지 점수에 집착하기 전에 반드시 알아야 할 PageSpeed 측정 가이드입니다.</description>
      <pubDate>Wed, 04 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/pagespeed-score-measurement-guide</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>SEO</category>
      <category>PageSpeed</category>
      <category>Lighthouse</category>
      <category>Core Web Vitals</category>
      <category>웹성능측정</category>
      <category>SEO</category>
      <category>LCP</category>
    </item>
    <item>
      <title>CPU 터보 부스트 끄기 — 서버 안정성을 위한 선택</title>
      <link>https://treeru.com/blog/cpu-turbo-boost-disable-server-stability</link>
      <description>Intel N100, AMD 5825U, 7840HS CPU의 터보 부스트를 끄는 방법과 서버 안정성에 미치는 실제 효과를 측정합니다. 온도 최대 37°C 감소, 전력 50% 절감, p99 지연 42% 개선까지 상시 가동 서버의 안정화 효과를 실측 데이터로 검증합니다.</description>
      <pubDate>Tue, 03 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/cpu-turbo-boost-disable-server-stability</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>하드웨어</category>
      <category>CPU</category>
      <category>터보부스트</category>
      <category>서버안정성</category>
      <category>Linux</category>
      <category>전력관리</category>
    </item>
    <item>
      <title>로컬 LLM 동시 접속 부하 테스트 - RTX PRO 6000으로 몇 명까지 처리할까?</title>
      <link>https://treeru.com/blog/local-llm-concurrent-users-load-test</link>
      <description>RTX PRO 6000에서 8B와 32B 모델로 동시 200명까지 부하 테스트한 실측 데이터. 응답 시간, P50/P95/P99 지연, GPU 온도, 처리량, VRAM 사용량, 에러율 0%까지 프로덕션 배포 전 반드시 확인해야 할 아키텍처 권장안과 최적 설정을 공개합니다.</description>
      <pubDate>Tue, 03 Feb 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/local-llm-concurrent-users-load-test</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>부하테스트</category>
      <category>동시접속</category>
      <category>RTX PRO 6000</category>
      <category>로컬AI</category>
      <category>SGLang</category>
      <category>프로덕션</category>
    </item>
    <item>
      <title>Tailwind CSS 애니메이션 라이브러리의 함정 — tw-animate-css 15KB 줄이기</title>
      <link>https://treeru.com/blog/tailwind-animate-css-bundle-size-optimization</link>
      <description>tw-animate-css를 전체 import하면 15KB CSS가 추가됩니다. 실제 사용하는 유틸리티만 추출해 3KB 이하로 줄인 실전 과정과 @keyframes 직접 작성법, Tailwind v4 @theme 설정, 렌더 블로킹 CSS 97% 감소까지 번들 최적화 기법을 공개합니다.</description>
      <pubDate>Sat, 31 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/tailwind-animate-css-bundle-size-optimization</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>Tailwind CSS</category>
      <category>tw-animate-css</category>
      <category>CSS번들</category>
      <category>PageSpeed</category>
      <category>웹성능최적화</category>
      <category>CSS최적화</category>
    </item>
    <item>
      <title>로컬 LLM 실무 테스트 (하) - 쇼핑/법률/자동화</title>
      <link>https://treeru.com/blog/local-llm-business-test-2</link>
      <description>6개 로컬 LLM으로 쇼핑몰 고객 응대, 법률 자문, 업무 자동화 3가지 시나리오를 직접 테스트합니다. 법률 조문 날조, 환불 정책 정확성, 윤리적 거부 패턴 등 실무 핵심 포인트를 모델별로 비교하고, 실제 서비스 투입 전 반드시 확인해야 할 위험 요소와 모델 선택 기준을 정리합니다.</description>
      <pubDate>Fri, 30 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/local-llm-business-test-2</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>LLM</category>
      <category>실무테스트</category>
      <category>법률AI</category>
      <category>쇼핑몰AI</category>
      <category>업무자동화</category>
      <category>로컬AI</category>
    </item>
    <item>
      <title>IndexNow로 검색 엔진에 즉시 알리기 — 크롤링을 기다리지 마세요</title>
      <link>https://treeru.com/blog/indexnow-instant-search-engine-notification</link>
      <description>블로그 발행 후 검색 노출까지 며칠을 기다릴 필요 없습니다. IndexNow 프로토콜 작동 원리와 Bing·Yandex·Naver 즉시 색인 요청 방법, API 구현 코드, 키 파일 관리, 제출 이력 추적, Next.js 자동화 통합까지 한 번에 정리한 SEO 실전 가이드입니다.</description>
      <pubDate>Wed, 28 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/indexnow-instant-search-engine-notification</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>SEO</category>
      <category>IndexNow</category>
      <category>SEO</category>
      <category>검색엔진</category>
      <category>Bing</category>
      <category>인덱싱</category>
      <category>크롤링</category>
    </item>
    <item>
      <title>로컬 LLM 실무 테스트 (상) - 제조/SaaS/의료</title>
      <link>https://treeru.com/blog/local-llm-business-test-1</link>
      <description>6개 로컬 LLM으로 제조업 부품 유통, SaaS 고객지원, 의료 상담 3가지 실무 시나리오를 30문항으로 테스트합니다. 각 모델의 응답 품질, 도메인 지식 정확성, 실제 서비스 투입 가능성을 실제 예시와 함께 비교하며 KORMo 1위 등 모델별 강점과 약점을 정리합니다.</description>
      <pubDate>Tue, 27 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/local-llm-business-test-1</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>LLM</category>
      <category>실무테스트</category>
      <category>제조업AI</category>
      <category>SaaS</category>
      <category>의료AI</category>
      <category>로컬AI</category>
    </item>
    <item>
      <title>Next.js &lt;img&gt; vs &lt;Image&gt; — PageSpeed 관점에서의 선택 기준</title>
      <link>https://treeru.com/blog/nextjs-img-vs-image-pagespeed</link>
      <description>Next.js Image 컴포넌트 자동 최적화와 img 태그 수동 최적화의 PageSpeed 점수 차이를 실측합니다. sharp AVIF 변환 후 img가 더 빠른 케이스, Image가 적합한 케이스, 각 방식의 빌드 시간·번들 크기·LCP 영향까지 실전 선택 기준을 정리합니다.</description>
      <pubDate>Tue, 27 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/nextjs-img-vs-image-pagespeed</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>Next.js</category>
      <category>Image컴포넌트</category>
      <category>img태그</category>
      <category>PageSpeed</category>
      <category>이미지최적화</category>
      <category>AVIF</category>
    </item>
    <item>
      <title>GPU 모니터링 자동화 — 셸 스크립트 하나로 24/7 감시</title>
      <link>https://treeru.com/blog/gpu-monitoring-shell-script-automation</link>
      <description>GPU 서버를 운영한다면 모니터링은 필수입니다. nvidia-smi와 cron을 조합해 5분 간격 GPU 로그를 CSV 포맷으로 수집하고, 온도·사용률 임계값 기반 알림, 로그 로테이션까지 50줄 내외의 셸 스크립트 하나로 해결하는 24/7 감시 자동화 실전 가이드입니다.</description>
      <pubDate>Mon, 26 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/gpu-monitoring-shell-script-automation</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>TOOL</category>
      <category>GPU모니터링</category>
      <category>nvidia-smi</category>
      <category>셸스크립트</category>
      <category>cron</category>
      <category>서버운영</category>
      <category>자동화</category>
    </item>
    <item>
      <title>리버스 프록시 서버 구성 — N100으로 충분한 이유</title>
      <link>https://treeru.com/blog/reverse-proxy-n100-mini-pc</link>
      <description>N100 미니PC에 Caddy 리버스 프록시를 구성해 자동 HTTPS, 멀티서비스 라우팅, 안정적 운영을 달성한 실전 과정. 2~5%의 낮은 CPU 사용률, Nginx와의 비교, Caddyfile 설정 예제와 보안 헤더, 장기 운영 안정성 데이터까지 저전력 홈서버 구축 가이드입니다.</description>
      <pubDate>Sat, 24 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/reverse-proxy-n100-mini-pc</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>네트워크</category>
      <category>리버스프록시</category>
      <category>Caddy</category>
      <category>N100</category>
      <category>미니PC</category>
      <category>HTTPS</category>
    </item>
    <item>
      <title>LCP 110초를 3.4초로 — motion/react의 opacity:0 트랩</title>
      <link>https://treeru.com/blog/lcp-optimization-motion-opacity-trap</link>
      <description>motion/react의 initial opacity:0이 SSR에서 이미지를 숨기고 LCP를 110초로 만드는 원리와, scale 전환·지연 로딩·AVIF 포맷 적용으로 3.4초까지 개선한 실전 과정. Framer Motion을 쓰면서 Core Web Vitals를 지키는 방법입니다.</description>
      <pubDate>Fri, 23 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/lcp-optimization-motion-opacity-trap</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>LCP</category>
      <category>motion/react</category>
      <category>Framer Motion</category>
      <category>PageSpeed</category>
      <category>Core Web Vitals</category>
      <category>SSR</category>
    </item>
    <item>
      <title>LLM 환각 테스트 - 거짓 정보를 만들어내는 모델은?</title>
      <link>https://treeru.com/blog/local-llm-hallucination-test</link>
      <description>6개 로컬 LLM에 존재하지 않는 제품, 가짜 판례, 의료 진단 유도 등 6가지 함정 질문을 던졌습니다. 어떤 모델도 전부 통과하지 못했고 Phi-4는 항생제 처방까지 시도하는 위험을 보였습니다. 환각 발생률과 치명적 결함, 모델별 안전 점수를 실제 응답과 함께 분석합니다.</description>
      <pubDate>Fri, 23 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/local-llm-hallucination-test</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>LLM</category>
      <category>환각</category>
      <category>Hallucination</category>
      <category>AI안전</category>
      <category>벤치마크</category>
      <category>로컬AI</category>
    </item>
    <item>
      <title>Drizzle ORM + SQLite로 가벼운 백엔드 만들기</title>
      <link>https://treeru.com/blog/drizzle-orm-sqlite-lightweight-backend</link>
      <description>PostgreSQL이 과한 프로젝트에서 SQLite가 답인 이유를 정리합니다. Drizzle ORM과 Prisma 비교, 타입 안전 스키마 정의, 마이그레이션 워크플로우, WAL 모드로 동시 읽기 성능 확보, 프로덕션 운영 팁까지 가벼운 백엔드 구축 실전 가이드입니다.</description>
      <pubDate>Tue, 20 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/drizzle-orm-sqlite-lightweight-backend</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>Drizzle ORM</category>
      <category>SQLite</category>
      <category>ORM</category>
      <category>백엔드</category>
      <category>데이터베이스</category>
      <category>TypeScript</category>
    </item>
    <item>
      <title>로컬 LLM 한국어 능력 비교 - 6개 모델 실전 테스트</title>
      <link>https://treeru.com/blog/local-llm-korean-test</link>
      <description>로컬 LLM 6종의 한국어 능력을 존댓말, 사자성어, 속담, 비즈니스 어투 등 10개 실전 질문으로 비교합니다. Qwen3·Gemma·KORMo·Llama·Phi-4 모델의 언어 오염, 자연스러움, 한국어 특화 테스트 결과와 실제 응답 예시까지 공개한 한국어 벤치마크입니다.</description>
      <pubDate>Tue, 20 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/local-llm-korean-test</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>LLM</category>
      <category>한국어</category>
      <category>로컬AI</category>
      <category>Qwen3</category>
      <category>KORMo</category>
      <category>Gemma</category>
    </item>
    <item>
      <title>이미지 최적화만으로 PageSpeed 20점 올리기 — 36MB를 0.8MB로</title>
      <link>https://treeru.com/blog/image-optimization-pagespeed-36mb-to-800kb</link>
      <description>Hero 이미지 6장(36MB)을 sharp로 0.8MB까지 압축한 실전 과정. WebP vs AVIF 포맷 비교, fetchPriority, lazy loading 전략, LCP 110초에서 6.8초 단축까지 PageSpeed 20점 이상 올리는 이미지 최적화 가이드를 공개합니다.</description>
      <pubDate>Mon, 19 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/image-optimization-pagespeed-36mb-to-800kb</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>이미지최적화</category>
      <category>sharp</category>
      <category>WebP</category>
      <category>AVIF</category>
      <category>PageSpeed</category>
      <category>LCP</category>
    </item>
    <item>
      <title>8B vs 14B vs 32B 동시 접속 비교 — 어떤 크기가 최적인가</title>
      <link>https://treeru.com/blog/local-llm-model-size-comparison</link>
      <description>RTX PRO 6000에서 Qwen3 8B·14B·32B 모델을 동시 200명까지 실측 비교합니다. 응답 속도 3배 차이, 품질 0.48점 차이의 트레이드오프, 처리량 스케일링, GPU 온도와 VRAM 사용량까지 모델 크기별 최적 선택을 위한 데이터를 종합 분석합니다.</description>
      <pubDate>Fri, 16 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/local-llm-model-size-comparison</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>AI</category>
      <category>모델크기비교</category>
      <category>8B</category>
      <category>14B</category>
      <category>32B</category>
      <category>동시접속</category>
      <category>로컬AI</category>
      <category>트레이드오프</category>
    </item>
    <item>
      <title>SSH 키 인증으로 다중 서버 관리하기 — 비밀번호 없는 안전한 접속</title>
      <link>https://treeru.com/blog/ssh-key-auth-multi-server-management</link>
      <description>SSH 키 인증으로 비밀번호 없이 여러 서버에 안전하게 접속하는 방법을 정리합니다. Ed25519 키 생성, 배포, 비밀번호 인증 비활성화, SSH config 별칭과 ProxyJump, fail2ban 설정, 보안 감사 로그까지 다중 서버를 안전하게 관리하는 실전 가이드입니다.</description>
      <pubDate>Fri, 16 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/ssh-key-auth-multi-server-management</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>네트워크</category>
      <category>SSH</category>
      <category>키인증</category>
      <category>서버관리</category>
      <category>보안</category>
      <category>Linux</category>
    </item>
    <item>
      <title>sharp 이미지 최적화 — 명령 하나로 웹 이미지 압축 자동화</title>
      <link>https://treeru.com/blog/sharp-image-optimization-automation</link>
      <description>Photoshop 없이 터미널에서 이미지를 최적화하는 개발자의 방법을 소개합니다. sharp 라이브러리로 WebP·AVIF 변환, 리사이즈, 200장을 30초에 처리하는 배치 스크립트, 품질 튜닝, Next.js 빌드 파이프라인 통합까지 이미지 최적화 자동화를 단계별로 정리합니다.</description>
      <pubDate>Thu, 15 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/sharp-image-optimization-automation</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>TOOL</category>
      <category>sharp</category>
      <category>이미지최적화</category>
      <category>WebP</category>
      <category>AVIF</category>
      <category>Node.js</category>
      <category>자동화</category>
    </item>
    <item>
      <title>Next.js 사이트 PageSpeed 최적화 — Mobile 38에서 88까지</title>
      <link>https://treeru.com/blog/nextjs-pagespeed-optimization-38-to-88</link>
      <description>Next.js 사이트의 PageSpeed Mobile 점수를 38에서 88까지 올린 실전 기록. 이미지 압축, CLS 제거, 한국어 웹폰트 최적화, LCP 개선, 렌더 차단 CSS 제거까지 4단계 과정과 각 단계별 측정 데이터를 공개한 Core Web Vitals 최적화 가이드입니다.</description>
      <pubDate>Wed, 14 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/nextjs-pagespeed-optimization-38-to-88</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>PageSpeed</category>
      <category>Next.js</category>
      <category>웹성능최적화</category>
      <category>Core Web Vitals</category>
      <category>LCP</category>
      <category>CLS</category>
    </item>
    <item>
      <title>RTX PRO 6000 토큰 속도 실측 - 350W에서 모델별 몇 tok/s?</title>
      <link>https://treeru.com/blog/rtx-pro-6000-llm-token-speed</link>
      <description>RTX PRO 6000 96GB에서 350W 전력 제한으로 6개 LLM의 토큰 생성 속도를 실측합니다. Qwen3-8B 208 tok/s부터 KORMo 60 tok/s까지 모델별 속도 측정과 속도-품질 트레이드오프, 실전 투입 권장 모델까지 GPU 활용 가이드를 상세히 분석했습니다.</description>
      <pubDate>Tue, 13 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/rtx-pro-6000-llm-token-speed</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>하드웨어</category>
      <category>LLM</category>
      <category>토큰속도</category>
      <category>RTX PRO 6000</category>
      <category>벤치마크</category>
      <category>SGLang</category>
      <category>AWQ</category>
    </item>
    <item>
      <title>Better Auth로 Next.js 인증 구현하기 — 세션 관리부터 관리자 권한까지</title>
      <link>https://treeru.com/blog/better-auth-nextjs-authentication</link>
      <description>NextAuth v5 마이그레이션이 부담스럽다면 Better Auth를 고려해보세요. 이메일·비밀번호 인증, 세션 쿠키 관리, 관리자 역할 분리, Drizzle ORM + SQLite 연동, 보안 강화까지 타입 안전한 Next.js 인증 시스템 구축 실전 가이드입니다.</description>
      <pubDate>Mon, 12 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/better-auth-nextjs-authentication</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>Better Auth</category>
      <category>Next.js</category>
      <category>인증</category>
      <category>세션관리</category>
      <category>관리자권한</category>
      <category>SQLite</category>
    </item>
    <item>
      <title>한국어 웹폰트가 PageSpeed를 깎는 이유 — 렌더 차단 CSS 26KB→3KB</title>
      <link>https://treeru.com/blog/korean-webfont-pagespeed-render-blocking-css</link>
      <description>next/font/google로 한국어 폰트를 로드하면 124개 @font-face가 생성되고 26KB 렌더 차단 CSS가 추가됩니다. 커스텀 서브셋과 font-display:optional, 셀프 호스팅으로 3KB까지 줄이고 CLS 0을 달성한 실전 최적화 과정을 공개합니다.</description>
      <pubDate>Sat, 10 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/korean-webfont-pagespeed-render-blocking-css</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>웹폰트</category>
      <category>한국어폰트</category>
      <category>PageSpeed</category>
      <category>렌더차단CSS</category>
      <category>font-display</category>
      <category>Next.js</category>
    </item>
    <item>
      <title>RTX PRO 6000으로 로컬 LLM 6종 벤치마크 - 종합 비교</title>
      <link>https://treeru.com/blog/rtx-pro-6000-local-llm-benchmark</link>
      <description>NVIDIA RTX PRO 6000 96GB에서 Qwen3, Gemma, Phi-4, Llama, KORMo 등 6개 로컬 LLM을 7개 시나리오 360개 질문으로 벤치마크한 종합 결과. 속도, 품질, 환각 트랩 통과율, 한국어 점수까지 모델별 강점과 약점을 실측 데이터로 비교합니다.</description>
      <pubDate>Fri, 09 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/rtx-pro-6000-local-llm-benchmark</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>하드웨어</category>
      <category>LLM</category>
      <category>벤치마크</category>
      <category>RTX PRO 6000</category>
      <category>로컬AI</category>
      <category>Qwen3</category>
      <category>Gemma</category>
    </item>
    <item>
      <title>iperf3로 사내 네트워크 속도 진단 — 1G vs 2.5G 실측</title>
      <link>https://treeru.com/blog/iperf3-office-network-speed-test</link>
      <description>iperf3로 사무실 LAN의 실제 속도를 정밀 측정하고 병목을 진단하는 실전 가이드. 1Gbps와 2.5Gbps NIC의 실측 차이, ethtool로 링크 속도 확인, 케이블·포트·허브 계층별 병목 원인 분석, 단계별 트러블슈팅 방법까지 네트워크 속도 문제 해결법을 정리합니다.</description>
      <pubDate>Thu, 08 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/iperf3-office-network-speed-test</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>네트워크</category>
      <category>iperf3</category>
      <category>네트워크속도</category>
      <category>LAN</category>
      <category>NIC</category>
      <category>네트워크진단</category>
    </item>
    <item>
      <title>CLS 0.5를 0으로 — TypeWriter 애니메이션의 함정과 해결</title>
      <link>https://treeru.com/blog/cls-zero-typewriter-animation-fix</link>
      <description>타이핑 애니메이션이 CLS 0.481을 만들고 있었습니다. Invisible Placeholder 패턴으로 시각적 변화 없이 CLS를 0으로 만들고 모바일 PageSpeed 점수를 25점 끌어올린 과정을 공개합니다. 애니메이션과 Core Web Vitals를 양립시키는 실전 기법입니다.</description>
      <pubDate>Tue, 06 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/cls-zero-typewriter-animation-fix</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>CLS</category>
      <category>Core Web Vitals</category>
      <category>PageSpeed</category>
      <category>애니메이션</category>
      <category>레이아웃시프트</category>
      <category>Next.js</category>
    </item>
    <item>
      <title>사무실에서 시작하는 AI 서버 인프라 — 16대 서버 실전 구성기</title>
      <link>https://treeru.com/blog/office-ai-server-infrastructure</link>
      <description>클라우드 없이 사무실에서 16대 서버로 AI 인프라를 구축한 실전 구성기. 서버 역할 분리, SSH 메시 보안, 2.5Gbps 네트워크 계층 설계, CPU 터보 안정화, 외부 서버 격리, 장기 운영 노하우까지 온프레미스 AI 인프라 구축의 전 과정을 상세히 공개합니다.</description>
      <pubDate>Tue, 06 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/office-ai-server-infrastructure</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>하드웨어</category>
      <category>서버인프라</category>
      <category>온프레미스</category>
      <category>SSH보안</category>
      <category>네트워크</category>
      <category>CPU안정화</category>
      <category>로컬AI</category>
    </item>
    <item>
      <title>AI 코딩 도구로 웹사이트 만들기 — 프롬프트만으로 어디까지?</title>
      <link>https://treeru.com/blog/ai-coding-tool-website-development</link>
      <description>AI 코딩 도구로 기업 홈페이지 4개를 직접 개발한 실전 경험을 공유합니다. AI 생성 코드 비율 65%, 개발 속도 3배 향상, 잘 되는 작업과 안 되는 작업의 명확한 경계, 효과적인 프롬프트 작성 팁, 비용 대비 생산성까지 솔직하게 분석한 1인 개발자의 현실 리뷰입니다.</description>
      <pubDate>Mon, 05 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/ai-coding-tool-website-development</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>TOOL</category>
      <category>AI코딩</category>
      <category>웹개발</category>
      <category>프롬프트</category>
      <category>생산성</category>
      <category>AI도구</category>
      <category>개발자</category>
    </item>
    <item>
      <title>웹사이트 배포 자동화 — Git Push부터 프로덕션 반영까지</title>
      <link>https://treeru.com/blog/web-deployment-automation-git-to-production</link>
      <description>CI/CD 도구 없이도 안정적으로 배포하는 방법을 공개합니다. Git push → build → PM2 restart 파이프라인 구축, 롤백 전략, 환경변수 관리, 다중 서버 동기화, 무중단 배포, 헬스체크까지 셸 스크립트만으로 완성한 실전 웹 배포 자동화 가이드입니다.</description>
      <pubDate>Sat, 03 Jan 2026 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/web-deployment-automation-git-to-production</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>배포자동화</category>
      <category>Git</category>
      <category>PM2</category>
      <category>CI/CD</category>
      <category>셸스크립트</category>
      <category>DevOps</category>
    </item>
    <item>
      <title>WireGuard VPN으로 외부에서 사무실 서버 접속하기</title>
      <link>https://treeru.com/blog/wireguard-vpn-office-remote-access</link>
      <description>OPNsense에 WireGuard VPN 서버를 구축해 외부에서 사무실 서버에 안전하게 접속하는 방법을 단계별로 설명합니다. 키 생성과 피어 설정, 노트북·iOS·Android 클라이언트 구성, 스플릿 터널링, 방화벽 룰, 연결 문제 트러블슈팅까지 실전 VPN 운영 가이드입니다.</description>
      <pubDate>Sat, 27 Dec 2025 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/wireguard-vpn-office-remote-access</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>네트워크</category>
      <category>WireGuard</category>
      <category>VPN</category>
      <category>원격접속</category>
      <category>OPNsense</category>
      <category>네트워크보안</category>
    </item>
    <item>
      <title>Caddy로 HTTPS 자동화 — Let&apos;s Encrypt 인증서 무한 갱신</title>
      <link>https://treeru.com/blog/caddy-https-auto-ssl-setup</link>
      <description>Nginx 설정에 지쳤다면 Caddy를 써보세요. Caddyfile 한 줄로 끝나는 리버스 프록시, 다중 도메인 운영, 보안 헤더, Let Encrypt 인증서 자동 갱신 설정까지 제로 컨피그 HTTPS 자동화를 구현하는 실전 가이드. Nginx 대비 설정 파일 90% 감소.</description>
      <pubDate>Tue, 23 Dec 2025 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/caddy-https-auto-ssl-setup</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>Caddy</category>
      <category>HTTPS</category>
      <category>Let&apos;s Encrypt</category>
      <category>SSL</category>
      <category>리버스프록시</category>
      <category>웹서버</category>
    </item>
    <item>
      <title>2024년 웹 디자인 트렌드 TOP 10</title>
      <link>https://treeru.com/blog/website-design-trends-2024</link>
      <description>2024년 웹 디자인 트렌드 TOP 10을 상세히 분석합니다. AI 기반 디자인, 다크모드, 글래스모피즘, 마이크로 인터랙션, 모노크롬, 3D 그래픽, 스크롤 스토리텔링, 가변 타이포그래피까지 실제 적용 사례와 구현 방법을 확인하고 트렌디한 웹사이트를 지금 만들어보세요.</description>
      <pubDate>Mon, 22 Dec 2025 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/website-design-trends-2024</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>디자인</category>
      <category>웹디자인</category>
      <category>트렌드</category>
      <category>UI/UX</category>
      <category>2024</category>
    </item>
    <item>
      <title>OPNsense로 사무실 네트워크 구축하기 — 공유기 대신 방화벽 라우터</title>
      <link>https://treeru.com/blog/opnsense-office-network-setup</link>
      <description>N100 미니PC에 OPNsense를 설치해 VLAN 세그먼트, 세분화된 방화벽 룰, WireGuard VPN까지 갖춘 소규모 사무실 네트워크를 300달러 이하로 구축한 실전 과정. 공유기 대신 방화벽 라우터로 기업급 보안을 확보하는 오픈소스 기반 네트워크 구축 가이드입니다.</description>
      <pubDate>Sat, 20 Dec 2025 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/opnsense-office-network-setup</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>네트워크</category>
      <category>OPNsense</category>
      <category>방화벽</category>
      <category>VLAN</category>
      <category>네트워크구축</category>
      <category>미니PC</category>
      <category>사무실네트워크</category>
    </item>
    <item>
      <title>Claude Code 실전 리뷰 — AI가 코드를 대신 짜주는 시대</title>
      <link>https://treeru.com/blog/claude-code-ai-coding-review</link>
      <description>Claude Code CLI 도구로 블로그 23편 작성, PageSpeed 38→88 최적화, 서버 관리 자동화를 3개월간 진행한 실전 리뷰. Cursor와의 비교, 장단점, 실제 비용 분석, 효과적인 프롬프트 패턴과 한계까지 AI 코딩 CLI의 현실을 솔직하게 정리합니다.</description>
      <pubDate>Fri, 19 Dec 2025 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/claude-code-ai-coding-review</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>TOOL</category>
      <category>Claude Code</category>
      <category>AI코딩</category>
      <category>CLI</category>
      <category>코딩도구</category>
      <category>Cursor</category>
      <category>AI개발</category>
    </item>
    <item>
      <title>웹사이트 SEO 최적화 완벽 가이드</title>
      <link>https://treeru.com/blog/seo-best-practices</link>
      <description>검색 엔진 최적화(SEO)를 위한 실전 가이드입니다. 메타 태그 작성법, 콘텐츠 전략, Core Web Vitals 최적화, 내부 링크, 외부 링크 빌딩, 기술 SEO, 지속적 모니터링까지 구글·네이버 상위 노출을 위한 모든 전략을 상세히 알려드리는 완벽 가이드입니다.</description>
      <pubDate>Thu, 18 Dec 2025 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/seo-best-practices</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>SEO</category>
      <category>SEO</category>
      <category>검색최적화</category>
      <category>마케팅</category>
      <category>구글</category>
    </item>
    <item>
      <title>PM2로 Next.js 프로덕션 운영하기 — 무중단 배포와 모니터링</title>
      <link>https://treeru.com/blog/pm2-nextjs-production-deployment</link>
      <description>Vercel 없이 자체 서버에서 Next.js를 운영하는 방법. PM2 ecosystem 설정, cluster 모드, reload 무중단 배포, pm2-logrotate, systemd 자동 재시작 연동까지 프로덕션 안정 운영에 필요한 모든 설정을 실전 예제와 함께 공개합니다.</description>
      <pubDate>Wed, 17 Dec 2025 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/pm2-nextjs-production-deployment</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>PM2</category>
      <category>Next.js</category>
      <category>무중단배포</category>
      <category>프로덕션</category>
      <category>서버운영</category>
      <category>Node.js</category>
    </item>
    <item>
      <title>반응형 웹 디자인의 중요성</title>
      <link>https://treeru.com/blog/responsive-web-design</link>
      <description>모바일 트래픽 60% 시대, 반응형 웹 디자인은 필수입니다. 유동적 그리드, 미디어 쿼리, 모바일 퍼스트 접근법, 유연한 이미지, 뷰포트 메타 태그, 브레이크포인트 설계까지 실제 코드 예시와 함께 반응형 웹 구현의 전 과정을 상세히 알려드리는 CSS 실전 가이드입니다.</description>
      <pubDate>Mon, 15 Dec 2025 00:00:00 GMT</pubDate>
      <guid isPermaLink="true">https://treeru.com/blog/responsive-web-design</guid>
      <author>admin@treeru.com (Treeru)</author>
      <dc:creator>Treeru</dc:creator>
      <category>개발</category>
      <category>반응형</category>
      <category>모바일</category>
      <category>CSS</category>
      <category>웹개발</category>
    </item>
  </channel>
</rss>