메뉴
HN
Hacker News • 5일 전

해적 얼굴, AI 모델 삭제 위기에서 구하다

IMP
7/10
핵심 요약

Pirate Face는 오픈소스 AI 모델(LLM, 이미지, 오디오, 데이터셋)을 토렌트·마그넷 링크로 변환해 P2P 방식으로 영구 보존하는 탈중앙화 인프라입니다. Hugging Face에서 삭제되더라도 파일 공유 무리(swarm)가 모델을 유지하며, 모든 파일은 Hugging Face 공식 SHA-256 체크섬으로 검증되어 변조를 방지합니다. 기존 파이프라인에서 환경변수만 바꾸면 되는 드롭인 API도 곧 제공될 예정입니다.

번역된 본문

AI를 영원히 살아있는 토렌트로 만드세요. 오픈 모델 - LLM, 이미지, 오디오, 데이터셋 - 을 절대 차단할 수 없는 🧲 마그넷 링크로 공개합니다. 단일 소유자나 단일 장애 지점이 없습니다.

pirateface.co/

위에 핸들을 입력하세요. 사용 가능 여부를 확인하려면 핸들을 입력하면 됩니다.

실시간 활동:

  • @zbce 핸들 등록 (최근)
  • @lynk 핸들 등록 (최근)
  • local-inference-lab/GLM-5.3-Flash-NVFP4 (최근)
  • Qwen/Qwen2.5-1.5B-Instruct (최근)
  • google/siglip2-base-patch16-256 (최근)
  • @frontierlabs 핸들 등록 (최근)
  • @frontierlabs 인증 완료 (최근)
  • tensorblock/openchat_3.5-gpt-4-80k-GGUF (최근)
  • @elxreno 핸들 등록 (최근)
  • @elxreno 인증 완료 (최근)
  • HuggingFaceTB/SmolVLM2-500M-Video-Instruct (최근)
  • catplusplus/Qwen3.8-27B-heretic-ara-NVFP4-MTP (최근)

전체 보기 →

인기 모델 Apache-2.0 및 MIT 라이선스 인기 모델을 Hugging Face에서 실시간 동기화하며, 각 모델은 체크섬 검증 토렌트가 됩니다.

  • prism-ml/Ternary-Bonsai-2-27B-gguf — 3일 전 업데이트 · 🧲 마그넷 · 텍스트 생성 · 54 GB · 🌱 5 시더 · 190만 조회
  • sentence-transformers/all-MiniLM-L6-v2 — 오늘 업데이트 · 🧲 마그넷 · 문장 유사도 · 91 MB · 🌱 18 시더 · 2억 5200만 조회
  • deepseek-ai/DeepSeek-V4.1-Flash — 10일 전 업데이트 · 🧲 마그넷 · 이미지→텍스트 · 765 GB · 🌱 7 시더 · 49만 7000 조회
  • convaiinnovations/laya — 어제 업데이트 · 🧲 마그넷 · 텍스트 분류 · 843 MB · 🌱 3 시더
  • XingChen-AGI/Xing4.0-29B-A4B — 2일 전 업데이트 · 🧲 마그넷 · 텍스트 생성 · 62 GB · 🌱 1 시더 · 1만 3000 조회
  • Qwen/Qwen3.8-27B — 37일 전 업데이트 · 🧲 마그넷 · 이미지→텍스트 · 56 GB · 🌱 1 시더 · 730만 조회
  • ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF — 18일 전 업데이트 · 🧲 마그넷 · 이미지→텍스트 · 12 GB · 🌱 2 시더 · 120만 조회
  • harshatheg/Qwen-2.5-1B-RLCD — 4일 전 업데이트 · 🧲 마그넷 · 텍스트 생성 · 준비 완료
  • baidu/Unlimited-OCR — 오늘 업데이트 · 🧲 마그넷 · 이미지→텍스트 · 6.7 GB · 🌱 6 시더 · 220만 조회

66만 9000개 이상의 대상 모델 탐색 →

주권 AI(Sovereign AI)를 위한 영구 보존 계층 Pirate Face는 주권 AI를 위한 탈중앙화 인프라입니다. 모든 오픈 모델은 Hugging Face에서 토렌트로 미러링되어 단일 회사가 아닌 P2P(피어 투 피어) 방식으로 보관됩니다.

모델은 절대 사라지면 안 되는 이유 →

검열 저항성 모든 모델은 Hugging Face에서 직접 다운로드할 수도 있는 토렌트입니다. Hugging Face에서 모델이 사라지는 날에도 파일 공유 무리(swarm)가 이를 유지합니다 — 차단할 수 있는 단일 호스트가 존재하지 않습니다.

huggingface.co/model이 삭제됨 → 폴백 ↓ pirateface.co/model — 1,240명 시딩 중

체크섬 검증 모든 파일에는 Hugging Face 공식 SHA-256 해시가 포함되어 있어 모든 바이트를 검증할 수 있습니다. 어디서 다운로드하든 해시가 일치해야 합니다 — 변조되지 않은 진짜 가중치만 받게 됩니다.

model-00001.safetensors — 6.7 GB SHA-256: 9f86d081884c7d659a2feaa0c55ad015a3bf4f1b2b0b822cd15d6c15b0f00a08 ✓ Hugging Face와 일치 — 비트 단위까지 변조 없음

드롭인 API 출시 예정 기존 파이프라인을 Pirate Face로 그대로指向하면 됩니다 — 같은 경로, 같은 API, 엔드포인트만 다릅니다. 코드 변경이 전혀 없습니다.

$ export HF_ENDPOINT=https://pirateface.co $ python train.py ✓ swarm에서 meta-llama/Llama-4를 가져오는 중

핸들 등록 pirateface.co/yourname을 공개 프로필로 예약하세요. 일치하는 Hugging Face 신원을 인증하면 '인증 크리에이터' 배지를 획득하고 사칭을 방지할 수 있습니다. 커뮤니티에 가입하면 무료 컴퓨팅 크레딧, 독점 모델 출시 등 향후 계정 혜택을 받을 수 있습니다.

계정 생성 — 계정이 없어도 탐색, 다운로드, 시딩은 가능합니다.

zai-org ✓ 인증됨 — pirateface.co/zai-org Z.ai · Zhipu AI 팔로우 - GLM 기반 모델, 베이징 LLM

  • zai-org/GLM-5.2 — 🌱 1,800 (LLM)
  • zai-org/GLM-4-9B — 🌱 1,240 (LLM)
  • zai-org/CogVideoX-5B — 🌱 760 (비디오)

자주 묻는 질문 Pirate Face란? 주권 AI를 위한 탈중앙화 P2P 계층입니다. (오픈 모...

원문 보기
원문 보기 (영어)
Turn AI into torrents that live forever. Open models - LLMs, image, audio, datasets - as 🧲 magnet links that can never be taken down. No single owner or point of failure. pirateface.co/ Enter your handle above Type your handle above to check availability. Live ⚑ @zbce claimed a handle · recent ⚑ @lynk claimed a handle · recent # local-inference-lab/GLM-5.3-Flash-NVFP4 · recent # Qwen/Qwen2.5-1.5B-Instruct · recent # google/siglip2-base-patch16-256 · recent ⚑ @frontierlabs claimed a handle · recent ✓ @frontierlabs verified · recent # tensorblock/openchat_3.5-gpt-4-80k-GGUF · recent ⚑ @elxreno claimed a handle · recent ✓ @elxreno verified · recent # HuggingFaceTB/SmolVLM2-500M-Video-Instruct · recent # catplusplus/Qwen3.8-27B-heretic-ara-NVFP4-MTP · recent ⚑ @zbce claimed a handle · recent ⚑ @lynk claimed a handle · recent # local-inference-lab/GLM-5.3-Flash-NVFP4 · recent # Qwen/Qwen2.5-1.5B-Instruct · recent # google/siglip2-base-patch16-256 · recent ⚑ @frontierlabs claimed a handle · recent ✓ @frontierlabs verified · recent # tensorblock/openchat_3.5-gpt-4-80k-GGUF · recent ⚑ @elxreno claimed a handle · recent ✓ @elxreno verified · recent # HuggingFaceTB/SmolVLM2-500M-Video-Instruct · recent # catplusplus/Qwen3.8-27B-heretic-ara-NVFP4-MTP · recent See all → Trending Models Trending Apache-2.0 & MIT models, synced live from Hugging Face - each becoming a checksum-verified torrent. prism-ml/Ternary-Bonsai-2-27B-gguf updated 3d ago 🧲 Magnet text-generation 54 GB 🌱 5 1.9M sentence-transformers/all-MiniLM-L6-v2 updated today 🧲 Magnet sentence-similarity 91 MB 🌱 18 252M deepseek-ai/DeepSeek-V4.1-Flash updated 10d ago 🧲 Magnet image-text-to-text 765 GB 🌱 7 497k convaiinnovations/laya updated yesterday 🧲 Magnet text-classification 843 MB 🌱 3 0 XingChen-AGI/Xing4.0-29B-A4B updated 2d ago 🧲 Magnet text-generation 62 GB 🌱 1 13k Qwen/Qwen3.8-27B updated 37d ago 🧲 Magnet image-text-to-text 56 GB 🌱 1 7.3M ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF updated 18d ago 🧲 Magnet image-text-to-text 12 GB 🌱 2 1.2M harshatheg/Qwen-2.5-1B-RLCD updated 4d ago 🧲 Magnet text-generation - Ready 0 baidu/Unlimited-OCR updated today 🧲 Magnet image-text-to-text 6.7 GB 🌱 6 2.2M Browse 669k+ eligible models → The permanence layer for sovereign AI. Pirate Face is decentralized infrastructure for sovereign AI. Every open model is mirrored from Hugging Face as a torrent, held peer-to-peer instead of by a single company. Why models should never die → Censorship-resistant Every model is a torrent that also downloads straight from Hugging Face. The day it's gone, the swarm keeps it alive - there's no single host to shut down. huggingface.co/model removed fallback ↓ pirateface.co/model 1,240 seeding Checksum-verified Every file carries its official Hugging Face SHA-256 - so you verify every byte. Download it anywhere and the hash still has to match: the real weights, never a tampered copy. model-00001.safetensors 6.7 GB SHA-256 9f86d0 81884c7d659a2feaa0c55ad015a3bf4f1b2b0b822cd15d6c15b0f00a08 ✓ Matches Hugging Face - bit-for-bit, untampered Drop-in API soon Point your existing pipeline at Pirate Face - same paths, same API, just a different endpoint. Zero code changes. pirateface - zsh # same pipeline - one env var $ export HF_ENDPOINT = https://pirateface.co $ python train.py ✓ pulling meta-llama/Llama-4 from the swarm Claim your handle Reserve pirateface.co/yourname for your public profile. Verify your matching Hugging Face identity to earn the Verified creator badge and help prevent impersonation. Join the community for upcoming account benefits, including free compute credits and exclusive model releases. Create account You can browse, download and seed without an account. zai-org ✓ verified pirateface.co/zai-org Follow Z.ai · Zhipu AI - GLM foundation models, Beijing LLM zai-org/GLM-5.2 🌱 1,800 LLM zai-org/GLM-4-9B 🌱 1,240 VIS zai-org/CogVideoX-5B 🌱 760 FAQ What is Pirate Face? A decentralized, peer-to-peer layer for sovereign AI. Open models from Hugging Face become checksum-verified torrents, held by a global swarm so they survive any single host taking them down. The goal is permanence for open-source AI. Do I need an account to use Pirate Face? No. You can browse, download, and seed without an account. Claiming is optional. A public claim reserves your Pirate Face handle and helps the network spread. Only your first eligible handle earns welcome points. Hugging Face creator claims require verification. To receive the Verified creator badge for a current Hugging Face username or organization, you must verify the matching Hugging Face account. Until then, the handle is only reserved and does not prove identity. Direct publishing is planned. Accounts will eventually support publishing and managing models directly on Pirate Face, without first uploading them to Hugging Face. That is not live yet. Today, submitting a model requires a Hugging Face account and the model must already be on Hugging Face. Accounts also support upcoming community benefits. See account benefits . Why create a Pirate Face account? An account gives you a place to manage your claimed handles and track your contributions. Claiming reserves a public handle; verifying the matching Hugging Face identity adds the creator badge and helps prevent impersonation. Pirate Face is preparing community benefits for account holders, including free compute credits and exclusive model releases . You can browse, download and seed without an account. Why use Hugging Face verification if the goal is independence from centralized hosting? Hugging Face verification confirms that someone claiming a creator’s handle controls the matching HF account or organization. It helps prevent impersonation and preserves attribution as models spread beyond their original host. You don’t need verification, or an account, to browse, download or seed. Pirate Face’s short-term goal is for model availability to outlive any single hosting platform. Today, parts of Pirate Face depend on Hugging Face and its community. Pirate Face is working to allow direct model publishing soon, without first uploading to HF. Does my Pirate Face handle have to match my Hugging Face username? No. You can reserve an available name, but the verified badge requires a matching Hugging Face identity. For example, a Hugging Face account named john123 can verify john123 , not john . You can still reserve john , but it stays unverified. The matching Hugging Face owner could reclaim that handle without deleting your Pirate Face account. Manage your connected accounts and reserved handles in Account . Could someone impersonate a creator here? Both are protected: The weights can't be faked - every file is checksum-verified against Hugging Face's official SHA-256. The handle - anyone can reserve a name (it shows reserved , no Verified creator ), but only proving the matching Hugging Face identity earns the Verified creator . A squatter can't lock you out: the real owner reclaims a reserved name by verifying. What does “checksum-verified” mean? Every weight file carries its official Hugging Face SHA-256 - a unique fingerprint of the exact bytes. Download it from anywhere and the hash still has to match, so you always know it's the real weights, never a tampered copy. It's the #1 fear with mirrored models, so it's a first-class feature here. How do downloads work, and what if Hugging Face removes a model? Every model is a magnet link (a torrent) with a Hugging Face web-seed built in - so it works even with zero peers. While it's on Hugging Face: you pull the bytes straight from HF - same weights, checksum-verified, same speed. The day HF removes it: the web-seed dies and the download falls back to the peer-to-peer swarm. We mark it Rescued - still reachable, kept alive by whoever's seeding. That fallback is the whole point: the same bytes as downloading from Hugging Face directly, but with no single point of failure (and