Magnific + GPT workflow: one detailed prompt for ultra-realistic paparazzi shots
techhalla · x · 2026-09-22
Creator techhalla shared a workflow for generating convincing candid paparazzi-style photos on Magnific: start with a GPT image model frame, attach a face reference photo, then guide the look with a detailed prompt.
The key trick is prompting for a "farther-distance phone photo" with smaller subjects and deliberately "mid quality — readable faces, mild phone softness," which avoids the overly polished AI look.
Full prompts were shared in the reply, making the faux-candid style reproducible.
More from Multimodal
- Grok Imagine demo shows 30-second continuous AI video generation — tetsuoai · 2026-09-22
- ElevenLabs launches Studio 4.0, an AI-native video editor with full model catalog built in — nikola_mr64990 · 2026-09-22
- Wyrdoom: an AI-generated gothic horror short film — EphemeralReverie_ · 2026-09-22
- An 8-chapter AI visual concept album follows one woman's entire life — Feeling-Total5959 · 2026-09-22
- Locally hosted H3 video model renders 25 seconds of 1942 Shanghai — CurieuxExplorer · 2026-09-22
- Open-source vectorizer InkVec beats VTracer, runs in under 2 seconds on CPU — incorporo · 2026-09-22