홈 / Skills / 이미지·영상 / 포스터·디자인 프롬프트 (Poster and Design Prompts)
이미지·영상

포스터·디자인 프롬프트 (Poster and Design Prompts)

AIBarsMITSkill 언어: 영어
라이선스 확인됨스크립트 없음낮은 위험
데모실행 기록 · claude-sonnet-5-5 · 2026년 10월
I'm making a poster for a neighbourhood jazz night at a small cafe called Blue Door Cafe. Details: Saturday 14 March, 7 pm, free entry, live trio. A3 portrait for printing and also a square version for social media. I'd like AI to make the whole poster in one go, including all the text, in a warm retro style. Can you write the prompt?
Skill: poster-design-prompts

Here are the prompts. You asked for the whole poster in one go, so the text is inside the image. AI tools sometimes garble lettering, so I kept the wording short, put every string in quotes, and left out the address (you didn't give one, and I won't invent it).

Prompt: A3 portrait

Vintage 1960s jazz club poster, vertical A3 portrait (aspect ratio 1:1.414, or the closest your tool offers). Warm retro screen-print style with visible paper grain, slightly worn ink texture and soft halftone shading. Limited palette: deep teal blue, burnt orange, mustard yellow, and cream paper, with a near-black for outlines.

Illustration: a stylised blue door in the center, slightly ajar and glowing with warm amber light. Silhouettes of a jazz trio (upright bass, piano, and drums or saxophone) are visible in the doorway. Musical notes and simple sunburst rays radiate outward.

Typography, spelled exactly as written, in bold retro display lettering with a clean sans-serif for the details:

  • Top, largest: "JAZZ NIGHT"
  • Directly beneath, medium: "at Blue Door Cafe"
  • Lower area, medium: "Saturday 14 March · 7 pm"
  • Bottom banner, smaller: "Live Trio · Free Entry"

Use these five lines of text only. Do not add any other words, URLs, addresses, logos or watermarks. Keep generous margins around all text, with high contrast between text and background. Centered, symmetrical layout with one clear focal point (the door) and a clear reading order from top to bottom.

Prompt: square (social media)

Vintage 1960s jazz club poster, square 1:1 format. Warm retro screen-print style with paper grain, worn ink texture and halftone shading. Palette: deep teal blue, burnt orange, mustard yellow, and cream, with near-black outlines.

Illustration: a glowing blue door slightly ajar, with a jazz trio in silhouette inside, and musical notes and sunburst rays around it. Keep the artwork compact so all text stays well inside the edges, because social platforms crop and overlay UI.

Typography, spelled exactly as written:

  • Top, largest: "JAZZ NIGHT"
  • Beneath: "at Blue Door Cafe"
  • Lower area: "Saturday 14 March · 7 pm"
  • Bottom banner: "Live Trio · Free Entry"

Use these four lines of text only. Do not add any other words, logos or watermarks. Centered layout, high contrast, one focal point.

Tips for getting a usable result

  • Generate several versions (4–8) and pick the one with correct text. Check every letter, especially "Saturday", "Cafe" and the "7 pm".
  • Generate the square separately. Don't crop the A3 version. The layout and text placement will differ.
  • If one line keeps coming out wrong: keep the best image and fix only that line in an editor (Canva, Photoshop, Affinity). If your tool supports inpainting, you can regenerate just that area.
  • If the text keeps failing: ask for the same artwork with "no text, no letters" and leave a calm area at the top and bottom. Then add the type yourself. That's the more reliable route, and I can write that prompt if you want it.

Print check for A3

  • A3 is 297 × 420 mm, which is about 3508 × 4961 px at 300 dpi. Most AI tools output less, so you'll likely need to upscale.
  • Ask your printer about bleed (usually 3 mm) and colour mode (usually CMYK). AI images are RGB, so colours, especially the teal, may shift slightly.
  • Keep important text well away from the edges.

Do you want to add an address or a cafe logo? If so, I'd add them in an editor afterwards. Small text and logos are the parts AI most often gets wrong.

Real run in an isolated folder with only this skill installed. Only the Skill and Read tools were enabled, so no command was run and no file was written. The scenario is fictional. No image was generated: the model only wrote text prompts, so how well they work depends on the image tool you use. Text rendering inside generated images is often unreliable; the answer says so and offers a typeset alternative. The print size figures were checked by hand afterwards.

하는 일

알려진 약점을 전제로 레이아웃 중심의 디자인을 계획합니다. 이미지 생성 도구는 이미지, 분위기, 색, 질감은 잘하지만 정확한 글자, 작은 글씨, 로고, 정렬은 불안정한 경우가 많기 때문입니다. 먼저 목적, 대상, 크기, 반드시 들어갈 문구가 담긴 브리프를 정리하고, 글자를 위한 빈 영역을 명시한 시각적 위계와 레이아웃을 정한 뒤, 하나의 스타일과 배색을 고르고, 작업을 나눕니다. 배경이나 일러스트용 프롬프트와, 디자인 프로그램용 글자 레이어 계획(단계, 크기, 정렬, 글꼴 조합)입니다.

글자에 대해 솔직합니다

이름, 날짜, 가격, 주소를 지어내지 않습니다. 그래도 글자를 이미지 안에 생성하고 싶다면 글자를 짧게 하고, 정확한 문자열을 따옴표로 묶고, 여러 번 시도할 것을 예상하게 하며, 글자 하나하나 교정하라고 알리고, 나중에 글자를 얹는 더 확실한 방법도 제시합니다.

출력

디자인 계획, 변형별 프롬프트, 자리표시자가 있는 글자 레이어 계획, 배색 메모, 체크리스트, 확인할 인쇄 사항.

참고 및 위험

낮은 위험:스크립트가 없는 지침 패키지로, 네트워크 접속이나 파일 쓰기가 없습니다. 프롬프트와 디자인 계획을 쓸 뿐 이미지를 생성하지 않습니다. 생성된 이미지 속 글자는 부정확한 경우가 많으므로 글자 하나하나 교정하고 날짜, 가격, 연락처는 직접 가진 원본 자료와 대조하세요. 글꼴, 로고, 이미지에는 적절한 라이선스가 필요하고 인쇄 사양(크기, 도련, 색상 모드)은 인쇄소에 확인해야 합니다. 기관, 공식 안내, 특정인을 사칭하는 자료나 가짜 초대장, 입장권, 증명서 제작은 돕지 않습니다. AIBars가 만든 오리지널(MIT). 가상의 카페 포스터로 한 번 시험 실행했으며 이미지는 생성하지 않았습니다.