A complete prompt library for the photography, video, and graphic identity of Estancia Cafetera.
Every image, every video, every graphic must answer one question: does this feel observed, or does this feel produced? The brand only ships the first.
Single-source natural light. The hour before sunrise, the hour after. Long shadows. Specular highlights on stone, ceramic, copper. No flash, no fill, no HDR. If the room is dim, photograph the dimness.
The closest thing to camera must show material truth — chisel marks on stone, throwing rings on ceramic, slubs on linen, crema on coffee. Never airbrushed, never staged.
Show the work, not the face. Barista hands tamping. Farmer hands sorting cherries. Cook hands kneading. Guest hands holding a cup. Faces only in candid, observed moments — never to camera.
Organized by surface use case. Forty hero images across eight categories. Generate in this order — landscape first to establish mood, lifestyle last because it depends on everything else.
The locked Midjourney v7 prompt structure. Every shot in Section 02 fills these tokens. Never deviate from the camera, palette, or negative blocks — those are what keep the library consistent.
/* MASTER PROMPT — Midjourney v7 syntax */ [SUBJECT] → filled per shot · see Section 04 [LIGHT] morning side light from left, raking, long shadows, 7am, window light, soft falloff, single light source, no fill [COMPOSITION] off-center subject, generous negative space, rule-of-thirds, low camera angle, asymmetric framing [MATERIAL] hand-thrown stoneware, dark wood, copper patina, slubbed natural linen, volcanic stone, visible texture, lived-in, unpolished, hand-touched [PALETTE] warm terracotta and deep brown earth tones, ash white linen, copper accent, deep green foliage, muted color, no saturation, no neon, no bright primaries [CAMERA] medium format film photography, Mamiya 7, Kodak Portra 400, slight grain, shallow depth of field, f/2.8, soft focus falloff [MOOD] quiet, observed, intimate, candid documentary, no performance, no posing, morning silence [NEGATIVE] --no people facing camera, smiling, logos, signage, modern technology, plastic, neon, HDR, drone view, wide grin, stock photography, tourist [FLAGS] --ar 4:5 /* portrait · default · for web hero */ --style raw /* less MJ stylization, more photographic */ --stylize 200 /* low-to-mid styling, prevents over-rendering */ --v 7
--ar 4:5 for Framer hero portraits, social verticals. --ar 3:2 for Framer wide blocks, room galleries. --ar 16:9 for video stills, header bars. --ar 1:1 for Instagram grid, thumbnail tiles.
--stylize 200 is the default. Going higher (400+) makes Midjourney add invented details — fake plants, ornamental serving ware, decorative kitsch. The brand fails. Never above 250.
Each prompt is complete and production-ready. Paste directly into Midjourney v7. No editing required.
Short atmospheric loops for hero motion, social, and the website's first scroll. Each clip 4–8 seconds. Camera movement is minimal: slow push, slow tilt, or static observation.
Brand graphics that live alongside the photography — signage, packaging, stamps, seals. Not photographic. Drawn, engraved, stamped. The Archivo Black wordmark lineage carries here.
Midjourney will give you four options per prompt. You'll be tempted to pick the prettiest. Don't. Pick the most true — the one that looks observed, not produced.
Generate in this order to build a consistent library, not a scattered one. Each phase establishes the visual language the next phase inherits. Don't skip ahead.
Generate Group 01 (6 prompts). This establishes the brand's atmospheric baseline — color temperature, light quality, grain. Every subsequent image references back to these.
Generate Group 02 (6 prompts). With the landscape mood locked, the interior shots can match. Stone, wood, copper, linen — the material palette becomes consistent.
Generate Group 07 (4 prompts). These tighter compositions inherit the palette from Phase 02 and become reusable across surface mockups, social, and the website's transition moments.
Generate Group 03 (6 prompts). The core brand story — from cherry to cup. These will be the most-used images in the library. Prioritize getting two or three exceptional ones over six average ones.
Generate Groups 04 + 05 (10 prompts). Mexican plates, breakfast scenes, drinks. Food is where Midjourney is most likely to invent decorative kitsch — curate hard, kill aggressively.
Generate Group 08 (4 prompts). Hands at work, no faces. v7 handles this well but watch for invented jewelry, watches, or distracting tattoos. Re-prompt if anything performs.
Generate Group 06 (4 prompts). These are the hardest because lifestyle is where AI most wants to add staged "wellness" energy. Lean toward empty hammocks and observed details, away from people-in-poses.
Once the still library is complete, move to Section 05 (Sora video prompts) and Section 06 (Kittl graphics). The videos extend the still moments; the graphics live alongside.