약관 및 정책

이용약관

서비스
Human.ai (휴먼ai)
시행일
2026년 11월 1일

Human.ai를 이용하기 전에 읽어 주시기 바랍니다. 이 약관은 서비스가 무엇이고 무엇이 아닌지, 이용자가 지켜야 할 것과 운영자가 책임지는 것을 정합니다.

법률 검토 전 초안

이 문서는 변호사의 검토를 마치기 전의 초안이며 확정본이 아닙니다. 점선으로 표시한 대괄호 부분은 운영자가 채워야 하는 정보이고, ‘검토 메모’는 검토하는 사람을 위한 안내입니다.

1.목적

이 약관은 크레이빙(CRAVING)(이하 ‘운영자’)이(가) 제공하는 Human.ai(휴먼ai, 이하 ‘서비스’)의 이용과 관련하여 운영자와 이용자의 권리, 의무와 책임에 관한 사항을 정하는 것을 목적으로 합니다.

2.운영자 정보

서비스를 운영하는 사업자

상호

크레이빙(CRAVING)

대표자

송현섭

주소

서울특별시 성북구 안암로 145, 경영본관동 2층 224호(안암동5가, 고려대학교안암캠퍼스)

전화

010-9027-0729

사업자등록번호

256-52-01047

통신판매업 신고

제2026-서울성북-1244호

호스팅서비스 제공자

Vercel Inc.

3.용어의 뜻

  • 서비스: 운영자가 웹사이트와 모바일 앱으로 제공하는 Human.ai의 심리 리포트와 이에 딸린 기능
  • 이용자: 이 약관에 따라 서비스를 이용하는 사람
  • 리포트: 이용자의 응답과 과제 수행, 또는 이용자가 제출한 콘텐츠를 바탕으로 서비스가 만들어 보여 주는 결과물
  • 성향 리포트: 연애 성향 리포트, 관계 안정감 리포트, 대인관계 스타일 리포트
  • 셀프체크: 마음 컨디션 체크, 집중 스타일 리포트
  • 인스타그램 리포트: 인스타그램 피드 리딩
  • 기본 리딩: 인스타그램 리포트를 무료로 받는 방식
  • 첫인상 리딩: 보낼 수 있는 게시물이 적을 때 무료로 받는 짧은 리딩. 자세한 내용은 요금과 유료 서비스에 적습니다.
  • 심층 리딩: 인스타그램 리포트를 앱에서 구매하여 받는 유료 방식. 자세한 내용은 요금과 유료 서비스에 적습니다.
  • AI 해석: 생성형 인공지능이 작성한 해석 문장
  • 기기: 이용자가 서비스를 이용하는 휴대전화, 컴퓨터 등의 단말기와 그 안의 브라우저 또는 앱

4.약관의 효력과 변경

  1. 운영자는 이 약관을 서비스 화면에서 언제든지 볼 수 있도록 게시합니다.
  2. 서비스에는 회원가입이 없습니다. 이 약관은 이용자가 서비스를 이용하는 동안 적용됩니다.
  3. 운영자는 관련 법령을 위반하지 않는 범위에서 이 약관을 변경할 수 있습니다. 변경하는 경우 변경 내용과 사유, 적용일을 적용일 전에 서비스 화면에 공지하며, 이용자에게 불리한 변경은 충분한 기간을 두고 미리 공지합니다.
  4. 변경된 약관에 동의하지 않는 이용자는 서비스 이용을 중단할 수 있습니다.

검토 메모

공지 기간(며칠 전에 알릴 것인지)과, 회원가입 없이 이용하는 서비스에 약관을 적용하는 방식이 「약관의 규제에 관한 법률」에 비추어 적정한지 검토가 필요합니다. 리서치 문서는 이 법의 내용을 확인되지 않은 기억으로 표시하고 있습니다.

5.서비스의 내용

서비스는 다음 리포트를 제공합니다. 실제로 제공하는 리포트의 종류는 서비스 화면에 표시된 바에 따릅니다. 스토어에서 내려받는 모바일 앱에는 셀프체크가 들어 있지 않습니다.

  • 인스타그램 피드 리딩
  • 연애 성향 리포트
  • 관계 안정감 리포트
  • 대인관계 스타일 리포트
  • 마음 컨디션 체크
  • 집중 스타일 리포트
  1. 서비스는 회원가입이나 로그인 없이 이용합니다.
  2. 서비스는 성인만 이용할 수 있습니다. 자세한 내용은 이용 연령에 적습니다.
  3. 문항 응답, 과제 기록, 그림, 완성된 리포트는 이용자의 기기에만 저장됩니다.
  4. 서비스는 무료로 이용할 수 있습니다. 인스타그램 리포트의 심층 리딩만 앱에서 구매하는 유료 서비스이며, 자세한 내용은 요금과 유료 서비스에 적습니다.

6.서비스의 성격과 한계

  1. 모든 리포트는 자기이해를 돕기 위한 참고 자료입니다. 의학적·심리학적 진단, 치료, 처방이 아니며 이를 대신하지 않습니다.
  2. 리포트는 이용자가 답한 내용이나 제출한 콘텐츠를 정리한 것입니다. 그날의 상태와 상황에 따라 달라질 수 있고, 맞지 않을 수 있습니다.
  3. 셀프체크에 관하여: 본 서비스는 의료기기가 아니므로 질병 유무를 판단할 수 없으며, 질병 유무 판단을 위해서는 의사의 검진이 필요합니다.
  4. 리포트를 채용, 진학, 법적 판단 등 중요한 결정의 근거로 사용하지 마시기 바랍니다.
  5. 서비스는 위기 상황에 대응하는 수단이 아닙니다. 마음이 많이 힘들거나 긴급한 도움이 필요한 경우에는 고객 지원의 ‘도움이 필요할 때’에 안내된 상담 전화나 긴급 전화로 연락하시기 바랍니다.

7.AI 해석에 관한 안내

  1. Human.ai의 AI 해석은 생성형 인공지능(AI)이 작성합니다. 사람이 직접 작성하거나 전문가가 개별적으로 검토한 결과가 아닙니다.
  2. AI 해석은 성향 리포트와 인스타그램 리포트에서만 제공합니다. 셀프체크에는 AI를 사용하지 않습니다.
  3. AI가 작성한 부분에는 AI가 생성하였다는 사실을 표시합니다.
  4. AI 해석은 이용자가 동의하고 요청한 경우에만 만들어지며, 이때 필요한 내용이 운영자의 서버를 거쳐 AI 서비스 제공자(Anthropic, PBC, 미국)에게 전송됩니다. 전송하는 내용과 처리 방식은 개인정보 처리방침에서 확인할 수 있습니다.
  5. 성향 리포트의 AI 해석에 대한 동의는 이용자의 기기에 기록되며 설정에서 언제든지 철회할 수 있습니다. 인스타그램 리포트의 동의는 리딩을 시작할 때마다 새로 받습니다.
  6. 성향 리포트는 AI 해석 없이도 척도 점수에 따른 기본 해석과 함께 이용할 수 있습니다.
  7. AI 해석은 사실과 다르거나 이용자에게 맞지 않을 수 있습니다.
  8. AI 서비스의 장애 등으로 AI 해석을 일시적으로 제공하지 못할 수 있습니다.

8.이용 연령

  1. 서비스는 성인만 이용할 수 있습니다. 대한민국에서는 만 19세 이상을 말합니다.
  2. 검사나 인스타그램 리포트를 처음 시작할 때, 그리고 AI 해석에 동의하기 전에 생년월일을 입력하여 연령을 확인합니다. 입력한 생년월일은 기기 안에서 나이를 계산하는 데만 쓰고 저장하거나 전송하지 않습니다.
  3. 이용자는 생년월일을 사실대로 입력하여야 합니다.

검토 메모

연령 확인은 이용자가 입력한 생년월일에 기대는 자기 확인 방식입니다. 본인확인 수단을 쓰지 않는 이 방식으로 성인 한정을 운영하여도 되는지, 스토어에 신고한 연령 등급과 타겟층에 맞는지 검토가 필요합니다. 미성년자가 사실과 다른 날짜를 입력하여 이용한 경우의 책임과 조치도 정하여야 합니다.

9.인스타그램 리포트의 이용

  1. 인스타그램 리포트는 이용자 본인의 계정만 읽습니다. 사용하는 자료는 이용자가 직접 넘겨준 것, 곧 이용자가 고른 본인 계정의 화면 캡처나 사진(피드 게시물, 스토리·하이라이트, 프로필 화면의 캡처 포함), 이용자가 인스타그램에서 내려받아 기기에서 읽어 들인 ‘내 정보 내보내기’ 파일, 또는(제공되는 경우, 웹에서만) 이용자의 프로페셔널 계정에 대한 Instagram 로그인 연결뿐입니다.
  2. Instagram 로그인 연결은 이용자가 직접 시작하고 Instagram의 화면에서 승인하여야 이루어집니다. 운영자는 Instagram 비밀번호를 받지 않으며, 이용자는 언제든지 연결을 해제할 수 있습니다.
  3. 계정 이름으로 다른 사람을 조회하는 기능은 없습니다. 운영자는 게시물을 자동으로 수집(스크래핑)하지 않으며, 그렇게 수집된 데이터를 구매하지도 않습니다.
  4. 이용자는 다른 사람의 계정이나 게시물을 분석할 목적으로 제출하여서는 안 됩니다.
  5. 이용자는 리딩을 시작할 때마다 제출하는 자료가 본인 계정의 것임을 확인하여야 합니다.
  6. 리포트는 피드가 보여 주는 모습을 정리한 것입니다. 해석마다 근거가 된 게시물을 붙이고, 관찰·경향·질문으로 구분하여 표시합니다. 첫인상 리딩은 고른 이미지가 보여 주는 첫인상만 정리하며, 여러 게시물에 걸친 경향은 말하지 않고, 관찰과 질문마다 근거가 된 이미지를 붙입니다.
  7. 리포트는 건강이나 정신건강 상태, 성적 지향, 인종·민족, 종교, 정치적 견해, 소득, 지능, 애착 유형, 미성년자에 관한 사항, 사진에 함께 나온 다른 사람에 관한 사항을 추정하지 않도록 설계되어 있습니다.
  8. 이용자는 리포트를 다른 사람에 대한 평가로 공개하거나 공유하여서는 안 됩니다.
  9. 제출한 콘텐츠에 대한 권리는 이용자에게 있습니다. 운영자는 리포트를 작성하여 제공하는 목적으로만 이를 처리하며, 운영자의 서버에 저장하거나 다른 목적으로 이용하지 않습니다. AI 서비스 제공자에게 전송되는 내용과 그 보유 기간은 개인정보 처리방침에 따릅니다.
  10. Human.ai는 Instagram 또는 Meta Platforms, Inc.와 제휴·후원·보증 관계가 없습니다.

검토 메모

이용자가 제출한 콘텐츠의 저작권 처리, 사진에 함께 나온 다른 사람의 취급, 이용자가 올린 ‘내 정보 내보내기’ 파일이 Meta 플랫폼 약관의 적용 범위 밖이라는 해석은 리서치 문서가 변호사 확인 사항으로 남긴 부분입니다. Instagram 로그인 연결은 Meta의 앱 검수를 통과하여야 제공할 수 있으며, 성격이나 성향을 평가하는 앱에 대한 Meta 개발자 정책 1.10의 적용 여부도 확인이 필요합니다.

10.이용자가 하여서는 안 되는 행위

  • 다른 사람의 계정, 게시물, 개인정보를 제출하는 행위
  • 리포트를 다른 사람을 평가하거나 비방하는 내용으로 공개하는 행위
  • 생년월일을 사실과 다르게 입력하여 연령 확인을 통과하는 행위
  • 자동화된 수단으로 대량의 요청을 보내거나 요청 제한 등 보호 조치를 우회하여 서비스의 정상적인 운영을 방해하는 행위
  • 서비스의 문항, 문구, 디자인, 소프트웨어를 운영자의 허락 없이 복제 또는 배포하거나 영리 목적으로 이용하는 행위
  • 그 밖에 법령을 위반하는 행위

운영자는 위 행위가 확인되면 해당 요청을 거부하거나 서비스 이용을 제한할 수 있습니다.

11.데이터의 보관과 이용자의 책임

  1. 리포트와 응답은 이용자의 기기에만 저장되므로, 운영자는 이를 복구하거나 다른 기기로 옮겨 줄 수 없습니다.
  2. 앱을 삭제하거나 브라우저에서 이 사이트의 데이터를 지우면 리포트와 응답도 함께 삭제됩니다. 보관이 필요하면 설정의 ‘내 데이터 내보내기’를 이용하시기 바랍니다.
  3. 내보낸 파일에는 리포트 내용이 그대로 담겨 있으므로 이용자가 안전하게 보관하여야 합니다.
  4. 모바일 앱의 리포트는 기기 백업에 포함되지 않도록 되어 있어, 기기를 바꾸거나 복원하면 따라오지 않습니다. 옮기려면 ‘내 데이터 내보내기’를 이용하시기 바랍니다.
  5. 기기를 다른 사람과 함께 쓰는 경우 기기 잠금 등으로 리포트를 보호하는 것은 이용자의 몫입니다.

12.지식재산권

  1. 서비스의 디자인, 문구, 소프트웨어 등에 대한 권리는 운영자 또는 정당한 권리자에게 있습니다.
  2. 서비스에 사용한 심리 척도는 각 출처가 정한 이용 조건에 따라 사용하며, 출처와 이용 조건을 해당 리포트에 표시합니다.
  3. 이용자는 자신의 리포트를 개인적인 용도로 보관하고 공유할 수 있습니다. 다만 인스타그램 리포트의 이용과 이용자가 하여서는 안 되는 행위에서 정한 제한을 지켜야 합니다.

13.요금과 유료 서비스

인스타그램 리포트에는 무료인 기본 리딩과 첫인상 리딩, 유료인 심층 리딩이 있습니다. 그 밖의 리포트와 기능은 무료입니다.

첫인상 리딩은 보낼 수 있는 게시물이 12개보다 적을 때(분류한 뒤 읽을 수 있는 게시물이 그보다 적은 경우 포함) 무료로 제공합니다. 이용자가 고른 본인 계정의 이미지(피드 게시물, 스토리·하이라이트, 프로필 화면의 캡처 등) 1장부터 11장까지를 Claude Opus 모델이 한 번의 요청으로 읽고, 보여 주기 전의 확인을 통과한 첫인상과 관찰 5개까지, 질문 3개까지를 근거가 된 이미지와 함께 보여 줍니다. 분류한 뒤에 제공하는 경우에는 분류 단계에서 어떤 해석의 근거로도 쓰지 않기로 한 게시물을 빼고 읽으며, 자해 표현이 감지된 경우에는 첫인상 리딩 대신 도움받을 수 있는 연락처를 보여 줍니다. 웹과 앱에서 이용할 수 있으며 결제가 없습니다.

구분기본 리딩심층 리딩
요금무료1회 13,800원. 실제로 청구되는 금액은 앱 마켓의 결제 화면에 표시된 금액입니다.
이용할 수 있는 곳웹과 앱앱 안에서만 구매할 수 있습니다. 웹사이트에서는 판매하지 않습니다.
필요한 게시물 수12개부터30개부터. 고른 게시물이 그보다 적으면 구매 화면에 심층 리딩을 보여 주지 않습니다.
AI 모델게시물의 분류와 해석 모두 Claude Haiku게시물의 분류와 해석 모두 Claude Opus
보여 주는 해석경향과 질문을 합하여 3개까지. 이를 잇는 설명 문단은 넣지 않습니다.보여 주기 전의 확인을 통과한 경향과 질문 모두와, 이를 잇는 설명 문단
  1. 결제: 심층 리딩은 모바일 앱 안에서 앱 마켓의 인앱 결제로만 구매할 수 있습니다. 결제는 앱을 내려받은 앱 마켓이 처리하며, 결제 수단과 결제 절차는 그 앱 마켓이 정한 바에 따릅니다. 운영자는 이용자의 이름, 이메일 주소, 카드 번호 같은 결제 수단의 정보를 받지 않습니다.
  2. 가격의 표시: 가격은 결제 전에 구매 화면과 앱 마켓의 결제 화면에 표시됩니다. 실제로 청구되는 금액은 앱 마켓의 결제 화면에 표시된 금액이며, 앱 마켓과 통화에 따라 위의 금액과 다를 수 있습니다.
  3. 한 번의 구매, 한 번의 리딩: 한 번 구매하면 한 번의 심층 리딩을 받을 수 있습니다. 리딩은 결제를 마치는 즉시, 이용자가 고른 게시물로 시작됩니다. 리포트가 만들어져 전달된 경우에만 그 구매를 사용된 것으로 봅니다.
  4. 분석할 수 있는 게시물이 30개보다 적은 경우: 심층 리딩에는 분석할 수 있는 게시물이 30개 이상 필요합니다. 게시물을 읽은 뒤 보호를 위해 따로 떼어 둔 게시물이나 읽지 못한 게시물이 있어 분석할 수 있는 게시물이 30개보다 적으면 심층 리딩을 시작하지 않습니다. 이 경우 그 구매는 사용된 것으로 보지 않으며 그대로 쓸 수 있고, 게시물을 더하여 다시 시도하거나 무료인 기본 리딩을 대신 받을 수 있습니다.
  5. 실패한 리딩: 리포트가 만들어지기 전에 리딩이 실패하면 그 구매는 사용된 것으로 보지 않으며, 추가 요금 없이 같은 구매로 다시 시도할 수 있습니다. 한 번의 구매로 리딩을 시작할 수 있는 횟수는 처음 시도를 포함하여 모두 8번까지입니다. 리딩에 앞서 게시물을 나누어 보내 분류하는 요청은 한 번의 구매로 모두 120번까지 받으며, 형식에 맞아 받아들여진 요청만 셉니다.
  6. 결과를 받지 못한 경우: 리딩이 성공한 뒤에도 연결이 끊겨 결과를 받지 못하는 경우에 대비하여, 첫 리딩이 성공한 때부터 1시간 동안 같은 구매로 두 번까지 더 요청할 수 있습니다. 이를 위해 앱은 리포트가 기기에 도착할 때까지 쓰던 구매를 기기에 남겨 둡니다.
  7. 리딩을 만들 수 없는 경우: 리포트가 전달되지 않은 채 위의 리딩 시도 횟수나 분류 요청 횟수를 모두 쓴 구매는 사용된 것으로 보지 않습니다. 앱은 그 구매를 닫지 않고 다음 리딩에 쓰지 않도록 따로 두며, 환불을 요청하도록 안내합니다. 환불은 구매한 앱 마켓의 환불 절차로 요청하거나 운영자에게 이메일이나 전화로 요청할 수 있으며, 방법은 환불 규정에 적습니다.
  8. 환불된 구매: 환불된 구매는 앱이 리딩에 쓰지 않으며, 심층 리딩을 새로 구매할 수 있습니다.
  9. 청약철회의 제한: 심층 리딩은 결제 즉시 제공이 시작되는 디지털콘텐츠입니다. 운영자는 결제 전에 구매 화면에서 리딩이 시작되면 청약철회가 제한된다는 점을 알리고 확인을 받으며, 무료인 기본 리딩으로 미리 이용해 볼 수 있게 합니다. 이에 따라 리딩이 시작된 뒤에는 「전자상거래 등에서의 소비자보호에 관한 법률」 제17조제2항제5호에 따라 청약철회가 제한됩니다.
  10. 청약철회와 환불의 조건, 기한, 방법은 환불 규정에 따릅니다.

검토 메모

판매를 시작하기 전에 확인할 사항입니다. 13,800원이 부가가치세를 포함한 금액인지 확인하여, 포함하지 않았다면 총액을 표시하여야 합니다(전자상거래법 제13조). 이 조항은 모델 이름을 적고 있으므로 운영 환경의 모델 설정(AI_MODEL_DEEP, AI_MODEL_FAST, AI_MODEL_LIGHT)을 바꾸면 함께 고쳐야 하며, 거절된 요청은 같은 제공자의 다른 모델이 대신 처리할 수 있다는 점(개인정보 처리방침의 ‘거절된 요청의 대체 처리’)도 이 표시와 함께 검토하여야 합니다. 구매하였으나 아직 쓰지 않은 심층 리딩(실패한 뒤 다시 시도할 수 있는 구매 포함)이 앱 삭제, 기기 변경, 서비스 종료 때 어떻게 되는지 정하여 이 조항과 서비스의 변경과 중단에 적어야 합니다. 그러한 구매는 앱 마켓이 끝나지 않은 구매로 보관하며, 앱 마켓에 따라서는 리딩을 전달하지 못한 구매를 며칠 뒤 자동으로 환불하므로(src/lib/billing/iap.ts), 다시 시도할 수 있는 기간을 어떻게 안내할지도 정하여야 합니다. 리포트가 전달되지 않은 채 횟수를 다 쓴 구매는 앱이 닫지 않고 환불을 안내하므로, 그 안내와 운영자의 환불 처리 절차가 환불 규정과 맞는지 확인하여야 합니다. 청약철회 제한의 문구는 환불 규정의 검토와 함께 변호사가 확인하여야 합니다.

14.서비스의 변경과 중단

  1. 운영자는 운영상 또는 기술상 필요에 따라 서비스의 전부 또는 일부를 변경하거나 중단할 수 있으며, 이 경우 미리 서비스 화면에 공지합니다. 미리 알릴 수 없는 부득이한 사유가 있는 경우에는 사후에 알릴 수 있습니다.
  2. 서비스를 종료하는 경우 운영자는 이용자가 그 전에 ‘내 데이터 내보내기’로 사본을 받을 수 있도록 미리 안내합니다.

15.책임의 범위

운영자는 운영자의 고의 또는 과실로 이용자에게 손해가 발생한 경우 관련 법령에 따라 책임을 집니다.

운영자는 다음의 경우 운영자에게 고의 또는 과실이 없는 한 책임을 지지 않습니다.

  • 천재지변 등 불가항력으로 서비스를 제공할 수 없는 경우
  • 기기의 고장이나 분실, 데이터 삭제 등 이용자의 기기에서 생긴 사유로 리포트가 없어진 경우
  • 이용자가 이 약관을 위반하여 서비스를 이용한 경우
  • 이용자가 리포트를 참고 자료의 범위를 넘어 진단이나 중요한 결정의 근거로 사용한 경우

검토 메모

책임 제한 조항은 「약관의 규제에 관한 법률」에 따른 검토가 필요합니다. 운영자의 책임을 포괄적으로 배제하는 문구는 넣지 않았습니다.

16.개인정보의 보호

운영자는 관련 법령에 따라 이용자의 개인정보를 보호합니다. 자세한 내용은 개인정보 처리방침에 따릅니다. 본인의 게시물이 다른 사람에 의해 제출되었다고 생각되는 경우에는 분석 제외 및 개인정보 요청 안내를 확인하시기 바랍니다.

17.문의와 분쟁의 해결

  1. 서비스에 관한 문의와 불만은 이메일 contact@craving.win 또는 전화 010-9027-0729(으)로 접수합니다. 운영자는 접수한 내용을 지체 없이 확인하여 처리합니다.
  2. 이 약관과 서비스의 이용에 관하여는 대한민국 법령을 적용합니다.
  3. 분쟁에 관한 소송은 관련 법령에서 정한 관할 법원에 제기합니다.
  4. 이용자는 한국소비자원, 콘텐츠분쟁조정위원회 등 분쟁 조정 기구에 조정을 신청할 수 있습니다.

검토 메모

준거법과 관할 조항, 국문본과 영문본이 다를 때의 우선 관계는 변호사가 정하여야 합니다. 초안에는 언어 우선 조항을 넣지 않았습니다.

18.시행일

이 약관은 2026년 11월 1일부터 적용합니다.

Policies

Terms of Service

Service
Human.ai (휴먼ai)
Effective date
1 November 2026

Please read these terms before you use Human.ai. They set out what the service is and is not, what you agree to when you use it, and what the operator is responsible for.

Draft pending legal review

This document is a draft that a lawyer has not yet reviewed, and it is not final. Bracketed text with a dashed outline marks information the operator still has to supply, and each “Review note” is guidance for the reviewer.

1.Purpose

These terms govern the rights, duties and responsibilities of 크레이빙(CRAVING) (the “operator”) and of you as a user in connection with Human.ai (휴먼ai, the “service”), which the operator provides.

2.About the operator

The business that runs the service

Company

크레이빙(CRAVING)

Representative

송현섭

Address

서울특별시 성북구 안암로 145, 경영본관동 2층 224호(안암동5가, 고려대학교안암캠퍼스)

Phone

010-9027-0729

Business registration no.

256-52-01047

Mail-order business registration

제2026-서울성북-1244호

Hosting service provider

Vercel Inc.

3.Definitions

  • Service: the Human.ai psychological reports and related functions that the operator provides through its website and mobile apps
  • User (“you”): a person who uses the service under these terms
  • Report: the result the service produces and shows you from your answers and task performance, or from content you submit
  • Style reports: Romantic style report, Relationship security report and Interpersonal style report
  • Self-checks: Mood check-in and Focus style report
  • Instagram report: Instagram feed reading
  • Basic reading: the free way to get the Instagram report
  • First look: a short free reading for when there are too few posts to send. The details are under Fees and the paid reading.
  • Deep reading: the paid way to get the Instagram report, bought in the app. The details are under Fees and the paid reading.
  • AI interpretation: interpretation text written by generative artificial intelligence
  • Device: the phone, computer or other terminal you use the service on, and the browser or app on it

4.Effect of these terms and changes to them

  1. The operator keeps these terms posted where you can read them in the service at any time.
  2. There is no sign-up. These terms apply for as long as you use the service.
  3. The operator may change these terms within the limits of the law. When it does, it will announce the change, the reason and the date it takes effect in the service before that date, and it will give a longer notice period for any change that is to your disadvantage.
  4. If you do not agree to changed terms, you may stop using the service.

Review note

The notice period (how many days in advance) and the way terms are applied to a service used without sign-up need review under Korea's Act on the Regulation of Terms and Conditions. The research brief marks its account of that Act as recalled, not verified.

5.What the service is

The service provides the reports below. Which reports are actually offered is as shown in the service. The mobile apps downloaded from the app stores do not contain the self-checks.

  • Instagram feed reading
  • Romantic style report
  • Relationship security report
  • Interpersonal style report
  • Mood check-in
  • Focus style report
  1. You use the service without signing up or signing in.
  2. The service is for adults only. The details are under Age requirements.
  3. Your answers, task records, drawings and finished reports are stored only on your device.
  4. The service is free to use. Only the deep reading of the Instagram report is a paid service, bought in the app; the details are under Fees and the paid reading.

6.What the service is not

  1. Every report is reference material for self-understanding. It is not a medical or psychological diagnosis, treatment or prescription, and does not replace one.
  2. A report summarises what you answered or the content you submitted. It can change with how you are and what is happening that day, and it may not fit you.
  3. About the self-checks: this service is not a medical device, so it cannot determine whether you have a disease. Determining that requires an examination by a doctor.
  4. Do not use a report as the basis for an important decision such as hiring, admissions or a legal judgment.
  5. The service is not a way to respond to a crisis. If you are struggling badly or need urgent help, contact one of the helplines or emergency numbers listed under “If you need help” on the Support page.

7.About AI interpretation

  1. AI interpretation in Human.ai is written by generative artificial intelligence (AI). It is not written by a person and is not individually reviewed by a professional.
  2. AI interpretation is offered only in the style reports and the Instagram report. The self-checks do not use AI.
  3. Parts written by AI are labelled as AI-generated.
  4. An AI interpretation is written only when you consent and ask for one. What it needs is then sent through the operator's server to the AI service provider (Anthropic, PBC, United States). What is sent and how it is handled is set out in the Privacy Policy.
  5. Your consent to AI interpretation of the style reports is recorded on your device, and you can withdraw it at any time in Settings. Consent for the Instagram report is asked afresh each time you start a reading.
  6. You can use the style reports without AI interpretation, with the built-in interpretation that follows from your scale scores.
  7. An AI interpretation can be wrong or may not fit you.
  8. AI interpretation may be temporarily unavailable, for example when the AI service has a fault.

8.Age requirements

  1. The service is for adults only. In Korea that means 19 or over.
  2. Your age is checked by asking for your date of birth the first time you start a test or the Instagram report, and before you consent to AI interpretation. The date you enter is used on your device only to work out your age and is neither stored nor sent.
  3. You must enter your date of birth truthfully.

Review note

The age check relies on the date of birth the user enters: a self-declaration. Whether an adults-only service may be run on this basis, without an identity check, and whether it matches the age rating and target audience declared to the app stores, needs review. What follows when a minor uses the service by entering a false date, and who is responsible, must be settled too.

9.Using the Instagram report

  1. The Instagram report reads your own account only. It uses only material you hand over yourself: screenshots or photos you pick from your own account (feed posts, stories and highlights, and captures of your profile screen included), your own “Export your information” file downloaded from Instagram and read on your device, or, where offered and on the web only, an Instagram Login connection to your own professional account.
  2. An Instagram Login connection happens only when you start it yourself and approve it on Instagram's own screen. The operator never receives your Instagram password, and you can disconnect at any time.
  3. There is no function for looking someone up by username. The operator does not collect posts automatically (scrape) and does not buy scraped data.
  4. You must not submit another person's account or posts for analysis.
  5. Each time you start a reading, you must confirm that the material you submit is from your own account.
  6. The report describes the picture your feed presents. Each interpretation is tied to the posts it rests on and is marked as an observation, a pattern or a question. A first look describes only the first impression the chosen images give: it states no tendency across posts, and each observation and question is tied to the images it rests on.
  7. The report is designed not to infer health or mental-health conditions, sexual orientation, race or ethnicity, religion, political views, income, intelligence, attachment style, anything about minors, or anything about other people who appear in your photos.
  8. You must not publish or share a report as an assessment of another person.
  9. You keep the rights in the content you submit. The operator processes it only to write and provide your report, and does not store it on its server or use it for any other purpose. What is sent to the AI service provider, and how long the provider holds it, is set out in the Privacy Policy.
  10. Human.ai is not affiliated with, sponsored by, or endorsed by Instagram or Meta Platforms, Inc.

Review note

The research brief leaves three points here for a lawyer: copyright in content users submit, the treatment of other people who appear in photos, and the view that a user-uploaded export file falls outside Meta's Platform Terms. The Instagram Login connection can be offered only after Meta's app review, and whether Meta's Developer Policy 1.10 on apps that assess personality or traits applies to it also needs checking.

10.What you must not do

  • Submit another person's account, posts or personal information
  • Publish a report as an assessment of, or an attack on, another person
  • Enter a false date of birth to pass the age check
  • Disrupt the normal running of the service, for example by sending requests in bulk by automated means or by getting around rate limits or other protective measures
  • Copy or distribute the service's questions, text, design or software, or use them commercially, without the operator's permission
  • Do anything else that breaks the law

If the operator finds any of the above, it may refuse the request concerned or restrict use of the service.

11.Your data and your responsibility for it

  1. Reports and answers are stored only on your device, so the operator cannot recover them or move them to another device for you.
  2. If you uninstall the app or clear this site's data in your browser, your reports and answers are deleted with it. To keep a copy, use “Export my data” in Settings.
  3. An exported file contains your full reports, so keep it somewhere safe.
  4. Reports in the mobile apps are set to stay out of device backups, so they do not follow you when you change or restore a device. To move them, use “Export my data”.
  5. If you share your device with others, protecting your reports, for example with a device lock, is up to you.

12.Intellectual property

  1. Rights in the service's design, text, software and so on belong to the operator or to their rightful owners.
  2. The psychological scales used in the service are used under the terms set by their sources, and each report states the source and those terms.
  3. You may keep and share your own reports for personal use, within the limits set under Using the Instagram report and What you must not do.

13.Fees and the paid reading

The Instagram report comes as a free basic reading, a free first look and a paid deep reading. Every other report and function is free.

The first look is free and is offered when you have fewer than 12 posts that can be sent (including when fewer than that could be read after coding). The Claude Opus model reads 1 to 11 images you chose from your own account (feed posts, stories and highlights, captures of your profile screen and the like) in one request, and shows a first impression, at most five observations and at most three questions that pass the checks made before anything is shown, each with the images it rests on. When it is offered after coding, posts the coding step kept from every interpretation are left out, and if words about self-harm were detected, help contacts are shown instead of a first look. It is available on the web and in the apps, and involves no payment.

AspectBasic readingDeep reading
PriceFree13,800 won per reading. The amount charged is the one shown on the app store's payment screen.
Where it is availableOn the web and in the appsBought inside the app only. The website does not sell it.
Posts neededFrom 12From 30. With fewer chosen posts, the purchase screen does not offer the deep reading.
AI modelClaude Haiku, for both the coding of your posts and the interpretationClaude Opus, for both the coding of your posts and the interpretation
Interpretation shownAt most three patterns and questions in all, without the paragraphs that connect themEvery pattern and question that passes the checks made before anything is shown, with the paragraphs that connect them
  1. Payment: a deep reading can be bought only by in-app purchase through the app store, inside the mobile app. The payment is handled by the app store you downloaded the app from, and the means and steps of payment are those that store sets. The operator does not receive your name, email address or payment details such as a card number.
  2. How the price is shown: the price is shown on the purchase screen and on the app store's payment screen before you pay. The amount charged is the one on the app store's payment screen, and it can differ from the amount above by store and currency.
  3. One purchase, one reading: one purchase gives you one deep reading. The reading starts as soon as payment is complete, on the posts you chose. A purchase counts as used only once its report has been made and delivered.
  4. Fewer than 30 posts that can be analysed: a deep reading needs at least 30 posts that can be analysed. If, once the posts have been read, fewer than 30 can be analysed because some were set aside for your protection or could not be read, the deep reading is not started. The purchase is then not treated as used and stays available: you can add posts and try again, or take the free basic reading instead.
  5. A reading that fails: if a reading fails before a report is made, the purchase is not treated as used, and you can try again with the same purchase at no extra charge. One purchase can start a reading up to eight times in all, the first attempt included. Before the reading, your posts are sent in parts to be coded; one purchase covers up to 120 of these coding requests in all, and only requests accepted as correctly formed are counted.
  6. If you do not receive it: in case the connection drops and you do not receive a reading that succeeded, the same purchase can be used to ask again up to two more times within one hour of the first successful reading. For this, the app keeps the purchase it was using on your device until the report arrives.
  7. If a reading cannot be made: a purchase that has used up the reading attempts or the coding requests above without a report being delivered is not treated as used. The app does not close it, keeps it aside so that it is not used for the next reading, and tells you to ask for a refund. You can ask through the refund process of the app store you bought it from, or ask the operator by email or phone, as set out in the Refund Policy.
  8. A refunded purchase: the app does not use a refunded purchase for a reading, and you can buy a new deep reading.
  9. Limits on withdrawal: a deep reading is digital content whose provision starts as soon as you pay. Before you pay, the operator tells you on the purchase screen that withdrawal is restricted once the reading starts and asks you to confirm, and it lets you try a reading first with the free basic reading. Accordingly, once the reading has started, withdrawal is restricted under Article 17(2)5 of Korea's Act on the Consumer Protection in Electronic Commerce.
  10. The conditions, time limits and methods for withdrawal and refunds are set out in the Refund Policy.

Review note

To be settled before anything is sold. Confirm whether 13,800 won includes VAT; if it does not, the total must be shown (E-Commerce Act, Article 13). This article names models, so it must change with any change to the production model settings (AI_MODEL_DEEP, AI_MODEL_FAST, AI_MODEL_LIGHT), and the fact that a declined request can be handled by another model from the same provider (“Fallback for a declined request” in the Privacy Policy) needs review alongside this statement. What happens to a deep reading that was bought but not yet used (including a purchase that can still be retried after a failure) when the app is deleted, the device changes or the service closes must be decided and written into this article and Changes to and suspension of the service. The app store keeps such a purchase open as unfinished, and depending on the store a purchase whose reading was never delivered is refunded automatically after a few days (src/lib/billing/iap.ts), so how long a retry remains possible, and how to say so, must be decided too. A purchase that used up its attempts without a report being delivered is not closed by the app, which tells the person to ask for a refund, so check that this message and the operator's own refund handling match the Refund Policy. A lawyer must check the withdrawal wording together with the review of the Refund Policy.

14.Changes to and suspension of the service

  1. The operator may change or suspend all or part of the service where operational or technical needs require it, and will announce this in the service in advance. Where advance notice is not possible for unavoidable reasons, notice may be given afterwards.
  2. If the service is to be closed, the operator will tell you in advance so that you can take a copy with “Export my data” first.

15.Liability

Where you suffer loss through the operator's intent or negligence, the operator is liable as the law provides.

Unless the operator acted with intent or negligence, it is not liable in the following cases.

  • The service cannot be provided because of force majeure such as a natural disaster
  • A report is lost for a reason arising on your device, such as a fault, loss of the device or deletion of data
  • You used the service in breach of these terms
  • You used a report beyond its role as reference material, as the basis for a diagnosis or an important decision

Review note

The limitation clauses need review under the Act on the Regulation of Terms and Conditions. No blanket exclusion of the operator's liability has been included.

16.Privacy

The operator protects your personal information as the law requires. The details are in the Privacy Policy. If you believe someone else submitted your posts, see Opt-out and privacy requests.

17.Enquiries and disputes

  1. Send enquiries and complaints about the service by email to contact@craving.win or call 010-9027-0729. The operator will look into them and act without delay.
  2. These terms and your use of the service are governed by the laws of the Republic of Korea.
  3. Any lawsuit over a dispute is brought in the court that has jurisdiction under the applicable law.
  4. You may apply for mediation to a dispute mediation body such as the Korea Consumer Agency (한국소비자원) or the Content Dispute Resolution Committee (콘텐츠분쟁조정위원회).

Review note

A lawyer must settle the governing-law and jurisdiction clauses and which language version prevails if the Korean and English texts differ. The draft contains no language-precedence clause.

18.Effective date

These terms apply from 1 November 2026.