15 Seconds — Final Fantasy VII Style — AI Video from Neta Studio

15 Seconds — Final Fantasy VII Style — AI Video from Neta Studio

Watch 15 Seconds — Final Fantasy VII Style — an AI-generated 15-second cinematic: a young woman in an oversized hoodie sits at a glowing gaming desk at night and says “You're still here.” Built in Neta Studio from a single sentence, free in your browser, no download.

Final Video: 15 Seconds — Final Fantasy VII Style

15 Seconds — Final Fantasy VII Style is a 15-second cinematic night scene rendered in the semi-realistic CG look of a late-2000s Final Fantasy VII cutscene: glossy skin and cloth, dramatic cinematic lighting, a moody colour grade — not cel-shaded, not flat cartoon.

In a small dark bedroom lit only by two monitors and blue-purple RGB strips, a late-night streamer realises the stream is over and the chat is empty. She stops performing, leans slightly toward the webcam, and speaks one honest line to the only person still watching — “You're still here.” — then sits in the quiet as the glow dims a touch. One close shot, one slow push-in, one tired smile.

Watch it now: 15 Seconds — Final Fantasy VII Style →

The Video

1080×1920 vertical · 15 seconds · 24 fps — rendered with Seedance 2.0 (image-to-video) from the world's own character and scene reference frames, in a single pass.

Download the MP4 →

How It Was Made

Production Process Breakdown

Stage 1: World Buildingclaude-opus-5
The world starts from one sentence — “Good vibe to work with, moody late night streamer setup… an emo-style girl streamer in an oversized long-sleeve hoodie, sitting at a gaming desk lit mostly by monitor glow and RGB LED strips in purple and blue.” Neta Studio expands it into a locked world: one character (the late-night streamer), one location (the small dark bedroom desk), one palette (cool monitor glow plus blue-purple RGB) and one mood (late-night quiet).
Stage 2: Style Explorationgpt-image-2
Two competing looks were generated from the same locked subject to test the feel: a handcrafted stop-motion version with visible clay, fabric and wire, and a cel-shaded 3D toon version with hard cel shadow bands and flat stepped colour. The cel look was carried forward as the working direction.
Stage 3: Reference Framesgpt-image-2
Two identity-locking references are generated: a character sheet of the young woman — front, three-quarter, back and a face close-up, dark straight hair, oversized grey hoodie, large over-ear headphones — and a scene reference of the bedroom, showing the two glowing monitors, the RGB strips on the wall and the warm mug on the desk.
Stage 4: Key Art Pagegpt-image-2
A vertical “STILL HERE” presentation page is laid out like a premium streaming service title page — a full-bleed hero of her face in monitor glow, a title lock-up, a “One viewer is still awake” line, and a nine-still contact sheet drawn from the same world.
Stage 5: First Render — Cel-Shaded Cutseedance-2-0
The first 15-second cut is rendered cel-shaded from the character sheet and the scene reference, following the exact beat timeline. It reads correctly, but the flat toon planes lose the skin and cloth detail the scene needs, so the look is re-briefed.
Stage 6: Final Render — Final Fantasy VII CGseedance-2-0
The same beat timeline and the same references are re-rendered in a semi-realistic cinematic CG style — detailed human features, glossy skin and cloth, dramatic cinematic lighting, moody colour grade, a late-2000s JRPG cutscene finish. One close webcam shot holds on her face with a slow push-in over the final beat; the finished MP4 publishes instantly to a shareable link that plays in any browser.

Generation Prompt

15 seconds. Final Fantasy VII cinematic CG style, semi-realistic — detailed human features, glossy skin and cloth, dramatic cinematic lighting, moody color grade, a late-2000s JRPG cutscene look; not cel-shaded, not flat cartoon. Small dark bedroom at night lit by monitor glow and blue-purple RGB strips, late-night moody light.

Reference images: [character reference] the young woman, oversized grey hoodie, dark straight hair, large headphones, locked identity and look; [scene reference] the room, the desk, two glowing monitors, RGB strips, warm mug, locked palette and layout.

Event goal: the stream is over and the chat is empty; she stops performing and speaks one honest line to the only person still watching, then sits in the quiet.

Spatial blocking: she sits at the desk facing the camera, a close webcam shot, two monitors glowing behind her, RGB strips along the wall behind her, a warm mug on the desk in the lower corner. Camera at her eye line, close on her face.

Camera: steady close shot just below her eye line holding on her face; a slow push-in over the final beat.

Beat timeline:
0:00-0:03 — close webcam shot of her face facing the lens. The RGB strips behind her shift from blue to purple across the wall; she glances to one side at the monitor, tucks a strand of hair behind her ear, and brings her eyes back to the lens. Low room hum and PC fan.
0:03-0:09 — the shot holds on her face. She leans slightly toward the lens, eyes on camera, and says "You're still here." Her voice is low and steady; the room hum drops.
0:09-0:15 — the same frame, slow push-in. A small tired smile; she looks down at her hands on the desk, then lifts her eyes back to the lens. The glow dims a touch. A soft breath; the room settling.

Sound: her spoken line "You're still here"; otherwise sound effects only, no background music.

Hard constraints: render in the referenced Final Fantasy VII semi-realistic CG style, not cel-shaded; keep her face consistent with the character reference, one character, no extra limbs, no duplicate character, no hand clipping, no text or captions anywhere in the scene.

Asset Prompts

Character reference — the streamer, four locked viewsgpt-image-2
“Character reference sheet in cel-shaded 3D game style, clean white background. An emo-style young woman in her early twenties, pale skin, dark straight hair falling past her shoulders, oversized grey long-sleeve hoodie, large over-ear headphones, tired eyes. Four views arranged left to right, evenly spaced: full front view, three-quarter view, back view, and a close-up of her face. Hard cel shadow bands, clean geometric silhouette edges, flat stepped colors with purple and blue accent lighting. No text, no labels, no borders, no watermark.”
Scene reference — the bedroom gaming setupgpt-image-2
“Cel-shaded 3D game render, vertical. A small dark bedroom at night lit by the blue and purple glow of a gaming setup. A young woman in an oversized grey hoodie and large headphones sits at a desk, seen from behind at a slight three-quarter angle, facing two glowing monitors. RGB strips on the wall cast blue and purple light; warm mug on the desk; dim room with posters barely visible in the background. Hard cel shadow bands, clean flat-shaded planes, low texture, late-night moody atmosphere. No text, no captions.”
Style test — handcrafted stop-motiongpt-image-2
“Use handcrafted stop-motion animation. Establish physical-model forms with visible assembly and slight construction irregularity. Tactile silhouette edges shaped by real materials and handmade joins. Material-led color with practical-camera value behavior. Practical directional lighting, grounded contact shadow, and shallow physical depth. Finish with clay, fabric, wood, wire, paint, seams, fingerprints, and handcrafted surface variation. Subject: an emo-style girl streamer in an oversized long-sleeve hoodie, sitting at a gaming desk, wearing headphones, pallid skin and muted tones. Scene lit mostly by cool monitor glow and purple-and-blue RGB LED strips; dim late-night room behind, posters barely visible. Camera positioned slightly above and in front of her at a three-quarter angle, capturing both her face and the glowing screens behind her. Late-night moody atmosphere. No text, no captions, no typography.”
Style test — cel-shaded 3D toongpt-image-2
“Use cel-style 3D toon rendering. Establish modeled three-dimensional forms simplified into readable anime-like planes. Selective contour rendering and clean geometric silhouette edges. Flat-to-stepped local colors with controlled saturation. Hard toon-shadow bands supported by coherent three-dimensional lighting. Clean stylized materials with low texture noise. Preserve the requested subject with cel-shaded 3D game rendering. Subject: an emo-style girl streamer in an oversized long-sleeve hoodie, sitting at a gaming desk, wearing headphones. Scene lit mostly by cool monitor glow and purple-and-blue RGB LED strips; dim late-night room behind, posters barely visible. Camera positioned slightly above and in front of her at a three-quarter angle, capturing both her face and the glowing screens behind her. Late-night moody atmosphere. No text, no captions, no typography.”
Key art page — “STILL HERE” presentation boardgpt-image-2
“A tall vertical key art page for an upcoming series, laid out like a premium streaming service's title page — restrained, cinematic, mostly image. The page ground is a deep desaturated deep blue-black. All interface furniture — type, thin dividers, pill outlines, the score ring — is flat, crisp and minimal, laid over the photography. Every photograph on this page is a photoreal live-action film still in full colour — real skin texture, real fabric, film grain, true photographic depth of field. No illustration, no painting, no drawn or cartoon look inside any photograph. The world in every still is the same: a small dark bedroom at night lit only by the blue and purple glow of a gaming setup. A young woman in her early twenties, pale skin, dark straight hair falling past her shoulders, an oversized grey long-sleeve hoodie and large over-ear headphones, tired eyes — she is the same person in every image. The hero still begins at the very top edge of the page and fills the top three fifths of it: one large full-bleed photograph of the young woman from the chest up, close in frame, leaning slightly toward the camera and looking straight into the lens, her expression soft and tired, the blue-purple glow of a monitor spilling across her face, a single warm mug on the desk in the lower corner, given plenty of room to breathe, its lower edge fading smoothly into the page ground. There is no bar, no band, no strip and no block of flat colour above the photograph or across the top of the page — the photograph itself reaches the top edge and the very first thing at the top of the page is the image. Above the title, the hero photograph runs uninterrupted all the way to the top edge of the page: that entire upper area is bare photograph with nothing whatsoever laid over it. The title is the very first element on the page. Laid over the lower left of the hero, the title reads "STILL HERE" in A thin, letter-spaced modern sans-serif in clean capitals. It is completely flat — no bevel, no three-dimensional extrusion, no metallic sheen, no chrome, no drop shadow and no glow. Directly beneath the title, one small thin line of metadata reads "15s · First person POV · Moody cel-shade". Below the hero, a thin full-width divider. Under it, on the left, a small upright poster thumbnail card with rounded corners showing the same bedroom as a dim silhouette, the desk and monitors a dark block against a faint blue-purple haze. To the right of that thumbnail: one short bold headline reading "One viewer is still awake", and beneath it two short lines of small grey body text reading "After the stream ends, the chat is empty. She glances at the screen once, then faces the camera.". Beneath that sit two rounded outline pill buttons side by side, the first reading "Character study" and the second reading "Late-night". At the right end of that same row, a small refined circular score ring: a single thin even stroke with rounded ends, sweeping about three quarters of the way round in warm amber. The remaining quarter of the circle is completed by a pale frosted arc of the same thickness that is clearly lighter than the page behind it, as though it were milky glass over the ground. It is never darker than the page. The number "8.7" is set large and light in the centre of the ring. There is no word or label of any kind beneath or beside the ring. Below that row, with no section label and no heading of any kind, a tidy grid of nine small stills, three across and three down, narrow even gutters, slightly rounded corners. The nine are different moments from the same world and are deliberately not in story order: her hand wrapped around a cold mug on the desk; the RGB strips sliding from blue to purple along the wall; her small silhouette in the dark in front of two glowing monitors; her headphones resting on her shoulders; a close-up of her tired eyes with the screen light reflected in them; a monitor screen reduced to a soft pale wash of light with everything on it blurred out of focus; her hand reaching up to pull the headphones off; a low angle of the desk edge with the warm mug and a single keycap catching blue light; her face in profile against the blue glow with the hood up, just turning toward the lens. Beneath the grid, centred in generous empty space, one line of quoted dialogue reads "You're still here.", set in A clean, slightly condensed humanist serif. The page ends there — there is no bar, no panel, no strip and no block of any kind below that line, only clean empty page ground. Every piece of text listed above is spelled exactly as given and written once only. There is no other text anywhere on the page — no section labels, no cast list, no credits, no additional captions, no ratings from any real review company, no other brand names, no page numbers, no signature, no watermark, no gibberish lettering and no invented words.”
First render prompt — cel-shaded cutseedance-2-0
“15 seconds. Cel-shaded 3D game render, small dark bedroom at night lit by monitor glow and blue-purple RGB strips, late-night moody light. Reference images: [character reference] the young woman, oversized grey hoodie, dark straight hair, large headphones, locked identity and look; [scene reference] the room, the desk, two glowing monitors, RGB strips, warm mug, locked palette and layout. Event goal: the stream is over and the chat is empty; she stops performing and speaks one honest line to the only person still watching, then sits in the quiet. Spatial blocking: she sits at the desk facing the camera, a close webcam shot, two monitors glowing behind her, RGB strips along the wall behind her, a warm mug on the desk in the lower corner. Camera at her eye line, close on her face. Camera: steady close shot just below her eye line holding on her face; a slow push-in over the final beat. Beat timeline: 0:00-0:03 — close webcam shot of her face facing the lens. The RGB strips behind her shift from blue to purple across the wall; she glances to one side at the monitor, tucks a strand of hair behind her ear, and brings her eyes back to the lens. Low room hum and PC fan. 0:03-0:09 — the shot holds on her face. She leans slightly toward the lens, eyes on camera, and says "You're still here." Her voice is low and steady; the room hum drops. 0:09-0:15 — the same frame, slow push-in. A small tired smile; she looks down at her hands on the desk, then lifts her eyes back to the lens. The glow dims a touch. A soft breath; the room settling. Sound: her spoken line "You're still here"; otherwise sound effects only, no background music. Hard constraints: keep her face consistent with the character reference, one character, no extra limbs, no duplicate character, no hand clipping, no text or captions anywhere in the scene.”

Reference Frames

Why Choose Neta Studio

Professional — The cloud harness supports the full workflow from world building to cinematic video, all running in the cloud with no local setup. Stable access to the latest models: Claude Opus 5 · GPT Image 2 · Seedance 2.0

One-click generation — Officially tuned skills drive every step: world building, style tests, character and scene references, key art, video generation, and instant publishing to a shareable link.

Community & IP-first — Adapt any world or start fresh from a single sentence. Keep your frames, prompts and sprites — remix them into new videos anytime.

Studio Interior

Start Creating

This video was generated by Neta Studio AI from a single sentence. The world took minutes. The video took minutes. You can do it too — adapt this world or start fresh from any IP. No coding required, free to start.

×