01
Spatial Scene Understanding
The multimodal model reads architectural geometry, surface materials, lighting, and visible site constraints directly from camera uploads.
Next-Gen Generative MarTech Platform
Autonomous Signage & Spatial Branding Powered by Multimodal AI
Turn street storefront photos and rough creative briefs into manufacturing-grade CAD specifications, photorealistic 3D renders, and accurate quotations in seconds.
vinhandesign-engine-v1.json
Claude Multimodal Reasoning
// Step 1: Input Analysis
INPUT: Facade image (1920x1080) + Brief ("Modern coffee house")
DETECTED: Width 4.8m, Height 1.2m, Concrete substrate
// Step 2: Reasoning Output
RECOMMENDATION: Halo-lit 3D Acrylic Lettering + Aluminum Composite Panel
STRUCTURAL SAFETY: Wind-load standard compliant (Grade A)
Instant Bill of Materials (BOM)
Substrate: Alu 3mm PVDF
5.76 m²
Illumination: Samsung 12V LED IP68
140 modules
Estimated Production Lead Time
48 Hours
Confidence Score:
98.4%
Seconds
From photo and brief to a structured specification
CAD + BOM
JSON ready for laser cutting and procurement
Multimodal
Vision and language in a single reasoning pass
Private Beta
Ho Chi Minh City fabrication network
Capabilities
Automating the Signage Lifecycle from Concept to Fabrication
02
Automated CAD & BOM Export
Manual drafting is replaced by structured JSON that fabrication teams convert into laser-cutting vectors and production bills of materials.
03
Brand Copy & Compliance
The system proposes storefront copy, typography, and contrast checks, then flags regional advertising rules for a human reviewer before production.
Pipeline
Four steps from a street photo to a shop-floor file
Each stage returns a deterministic artifact. Designers and fabricators stay in the loop before anything is cut.
-
01 — Capture
Upload the site
A facade photo plus a short brief: brand, tone, and must-keep constraints.
-
02 — Interpret
Read the space
Claude vision estimates span, substrate, and mounting conditions from the image.
-
03 — Specify
Lock the spec
Lettering, illumination, materials, and a priced bill of materials in structured JSON.
-
04 — Produce
Hand off to the shop
Export vectors and a production packet. A fabricator approves before cutting.
Use Cases
Built for teams that repeat signage at scale
QSR and coffee rollouts
Keep halo-lit lettering, panel specs, and illumination consistent across new storefronts.
Retail chains
Turn a brand book and a site photo into a quote the local fabricator can actually build.
Real estate and property
Monument signs, wayfinding, and leasing identity with dimensions taken from the facade.
Corporate campuses
Directory, reception, and exterior identity packages with a shared material system.
Engineering Architecture
Built on Anthropic Claude Reasoning Models
Vinh An Design uses Claude vision and long-context reasoning to translate unstructured spatial cues into engineering schemas. Outputs stay structured so downstream CAD and quoting tools can consume them without a second interpretation step.
- → Image plus brief in, validated JSON specification out
- → Human approval before any fabrication file is released
- → REST and webhook delivery into existing shop systems
Claude Multimodal
Structured JSON Output
REST & Webhook APIs
10x
Faster turnaround from brief to a production packet
Early Access
Request a developer beta seat
We are onboarding signage studios, brand teams, and fabrication shops. Tell us what you produce and we will reply from the address below.
- Studio
- Ho Chi Minh City, Vietnam
- Domain
- vinhandesign.io.vn
- Status
- Private Beta