GPT writes a real story arc — hook, conflict, twist, payoff. Natural dialogue, bilingual CN + EN.
Generate films & 3D worlds
from a single sentence.
mcstory is a text-to-3D engine. One sentence becomes a review-gated narrative video draft, or a downloadable 3D model candidate: buildings, characters, creatures, whole worlds. Proven inside Minecraft, the biggest game on Earth, and built to generalize with review gates.
One engine. Two products.
Everything here runs on one text-to-3D engine. It powers two products — and they compose.
A · Text → Film
One sentence → a review-gated narrative video draft: script, scene, cast, camera, voiceover, music, subtitles. The widest use: creators, marketing, education, entertainment, IP shorts.
B · Text → Model
One sentence → any downloadable 3D model: buildings, characters, creatures, vehicles, whole worlds. For game dev, builders, education, digital assets.
COMING SOON
A + B · Compose
A fully-customizable product: generate a unique model, then build your film around it — one seamless pipeline. Launching soon.
What makes it different
Not a template machine. Each beta run can draft, build, cast and shoot from scratch — with outputs reviewed before any public use.
Real, not mockups
Gallery items are internal beta outputs from text prompts, selected only after review gates.
Unique every time
A generative engine builds a one-of-a-kind scene and cast for each story — no reused sets, no stock templates.
Review-gated output
HD shaders, per-character voiceover, music and SFX, burned-in bilingual subtitles. Exports are reviewed before sharing or publishing.
Create your film
One sentence in — the system writes, builds, directs, shoots and dubs the whole thing. Self-serve film generation is in private beta: preview the flow below, or request access to generate your own.
Guided run
Deep edit
Genre
Cozy tavern
Comedy
Heartfelt
Study tale
Grand feast
Family life
Length
Short 20s
Standard 35s
Full 60s
~3 min · 1080p · voiceover, subtitles, BGM included
ScriptBuild scene
Direct castCamera
DubDone
Product A · Films from one sentence
Each film is generated end to end from a single line — scene, cast, dialogue, camera, voiceover and subtitles.
How it works
One sentence in, a reviewable short-film draft out — six assisted steps
1
Text → Script
GPT expands one line into a script with arc, beats and dialogue
2
Script → Scene
A generative engine builds a unique, detailed interior for each story
3
Scene → Cast
Characters generated by role and placed in position — not a crowd of extras
4
Cast → Performance
Actions, gestures, reactions — the picture actually performs the script
5
Performance → Camera
Cinematic camera: wide for narration, cut to whoever's speaking
6
Camera → Film
Bilingual voiceover + music & SFX + subtitles — MP4 draft for review
Everything in one sentence
Type one line. The beta pipeline assists the production steps and keeps outputs reviewable.
Characters act in the world — gestures, dialogue, reactions. Not static dummies.
Cinematic camera — establishing shots, dialogue framing, dynamic moves synced to the beats.
Every story gets a unique built environment — interiors and outdoor villages, docks, forests.
A sentence into a real 3D build — preview, download the .schem, drop it into your own world.
Per-character voices, scene-matched SFX, and mood-driven BGM — mixed for review.
HD shaders, burned-in bilingual subtitles, 16:9 + 9:16 for every platform.
One sentence to a reviewable video draft — fewer manual steps, no MC skills needed.
Product B · Generate any 3D model
The same engine turns one sentence into a downloadable 3D model — and builds it block by block before your eyes.
One sentence → 3 AI reference views → the model builds itself
“a white marble palace with a great central dome and four minarets” — generated from text: 3 AI reference views, then a 3D model that builds itself
Generate your own build — then download it · Controlled beta
Pagoda
Pyramid
Temple
Palace
Every model below was generated from the one line under it — not a template, not a library asset.
Pricing
Credits power everything — 1 film = 10 credits · 1 generated 3D model = 4 credits. A model you generate can also be dropped into a film (you pay for each). Every plan includes monthly credits; top up anytime.
30 credits / mo (~3 films). Watermarked exports, scene & build tries, 7-day retention.
250 credits / mo (~25 films). No watermark, MP4/SRT/.schem, 9:16 + 16:9, 1 concurrent job.
900 credits / mo (~90 films). Script editing, manifests, priority queue, 3 concurrent jobs.
3,500 credits / mo (~350 films). Shared workspace, seats, pooled credits, review links, webhooks.
Dedicated workers, SSO, private storage & deployment, audit logs, contracted credit volume + SLA.
Out of monthly credits? Buy more anytime — $10 → 80 · $40 → 360 · $90 → 900. Top-up credits never expire.
Building on our API? It's a separate plan — see below.
Partner API — limited beta
For selected partners integrating mcstory into their own product. Separate from creator plans, priced on usage: $99/mo platform + $0.08 per credit (1 generated 3D model = 4 credits, 1 film = 10 credits), with volume discounts. Scoped keys, REST jobs, webhooks and signed URLs are available by arrangement. Scene/model access is partner-limited; video generation remains controlled beta.
POST /api/v1/runs
Authorization: Bearer mcs_live_***
{
"mode": "scene",
"prompt": "a grand Gothic cathedral with twin spires and a rose window",
"outputs": ["schem", "preview_png", "manifest"],
"webhook_url": "https://example.com/mcstory-webhook"
}
→ { "run_id": "run_20260605_abc123", "status": "queued", "status_url": "/api/v1/runs/run_20260605_abc123" }Auth + scoped keysQueue + status pollingrun_manifest.jsonSigned artifact URLsRate limits + monthly creditsWebhook callbacks
Get in touch
For partnerships, studio plans, partner API access or private cloud deployment, reach out:
napkinstudio@hotmail.com
mcstory is a NapkinStudio product