핵심 요구사항
- Seedance를 중심 파이프라인으로 한 초사실적 세로형 AI 영상 생성 및 컴포지팅
- 초반 3초 후킹·클리프행어 중심의 짧은 호흡 연출력
- 번인 자막(또는 SRT), 다이내믹 텍스트, 알고리즘 친화 메타데이터 처리
- 9:16 무레터박스·오디오 피킹 없는 틱톡 즉시 업로드용 마스터링
- 시즌 연속성 유지를 위한 스타일 가이드 작성
Required stack
필요 기술
SeedanceStable DiffusionMidjourneyRunwayAI Video GenerationSRT/Captioning
Project brief
프로젝트 내용
I’m producing a 30-episode vertical drama built for TikTok’s algorithm: two-minute chapters, 9:16 aspect ratio, ultra-realistic visuals, and seamless AI generation through Seedance. The core vibe is dark, gritty, and relentlessly intense—an action story layered with fantasy and thriller elements that centres on anti-heroes and outright villains rather than traditional saviours.
Every episode must hook viewers in the first three seconds, sustain momentum with sharp pacing, and close on a cliff-hanger that drives re-watchs and shares. High-fidelity graphics are non-negotiable; I’m looking for near-cinematic quality despite the short form. Dialogue, animated movement, lighting, and atmospheric sound design all need to feel as if they were shot on location in a big-budget production, only delivered through AI.
Seedance will be the production backbone, so you should already be comfortable prompting, refining, and compositing with it. If you like to mix in tools such as Stable Diffusion, Midjourney, or Runway, great—as long as Seedance remains the central pipeline. Final masters must be ready for immediate upload to TikTok, complete with hard-coded captions, dynamic text where relevant, and algorithm-friendly metadata.
Deliverables
• 30 vertical episodes, 2:00 each (±5 s)
• Seedance project files and any supporting assets
• SRT or burned-in captions matching dialogue
• One-sentence hook and hashtag set per episode
• Style guide so future seasons can maintain continuity
Acceptance criteria: photorealism holds up on a 6-inch screen; no black bars, pillarboxing, or audio peaking; pacing never drags; and Seedance source files open error-free.
If you can marry storytelling flair with airtight AI production skills, let’s talk.
Every episode must hook viewers in the first three seconds, sustain momentum with sharp pacing, and close on a cliff-hanger that drives re-watchs and shares. High-fidelity graphics are non-negotiable; I’m looking for near-cinematic quality despite the short form. Dialogue, animated movement, lighting, and atmospheric sound design all need to feel as if they were shot on location in a big-budget production, only delivered through AI.
Seedance will be the production backbone, so you should already be comfortable prompting, refining, and compositing with it. If you like to mix in tools such as Stable Diffusion, Midjourney, or Runway, great—as long as Seedance remains the central pipeline. Final masters must be ready for immediate upload to TikTok, complete with hard-coded captions, dynamic text where relevant, and algorithm-friendly metadata.
Deliverables
• 30 vertical episodes, 2:00 each (±5 s)
• Seedance project files and any supporting assets
• SRT or burned-in captions matching dialogue
• One-sentence hook and hashtag set per episode
• Style guide so future seasons can maintain continuity
Acceptance criteria: photorealism holds up on a 6-inch screen; no black bars, pillarboxing, or audio peaking; pacing never drags; and Seedance source files open error-free.
If you can marry storytelling flair with airtight AI production skills, let’s talk.
지원 기회 분석
지원 전에 볼 것
예상 산출물
- 세로형 30편(각 2분±5초) 완성본
- Seedance 프로젝트 파일 및 보조 에셋
- 대사 일치 자막(SRT 또는 번인)
- 에피소드별 한 줄 훅 + 해시태그 세트
- 시즌 연속성용 스타일 가이드
매력 포인트
- 산출물과 인수 기준이 비교적 명확하게 명시됨
확인할 점
- 30편 초사실적 시네마틱 품질 요구 대비 예산($250~750)이 극단적으로 낮음
- 클라이언트 국가·평판 정보 부재(employer_id 0)
- 품질 기대치와 산출물 범위가 예산·기간에 비현실적
Client signal
클라이언트 정보
플랫폼에서 상세 정보를 확인하세요.