Prompt-Driven Building Plan Generator
Summary
Freelancer Client is hiring: Prompt-Driven Building Plan Generator.
Location: Remote
I need a full-stack solution that turns natural-language prompts into production-ready architectural documents. A user will visit the site, type a detailed description of the building they have in mind, specify every key dimension in plain text, press “Generate,” and immediately receive:
What you'll do:
• Downloadable outputs in at least DWG and IFC (plus a browser preview)
Nice to have:
• Supply a sample prompt library and the resulting DWG/IFC files for QA
• Pass an automated test set that checks dimensional accuracy and style compliance
• I will sign off when the test set passes and the deployment guide lets me replicate the stack on my own AWS account without modification.
Skills: CAD/CAM, Building Architecture, Home Design, AutoCAD, Machine Learning (ML), Full Stack Development, Vue.js, Architecture, REST API, Natural Language Processing
Budget: $600–$1000 USD
Source: Freelancer Client via Remote / Online. Apply on the source website.
Original
I need a full-stack solution that turns natural-language prompts into production-ready architectural documents. A user will visit the site, type a detailed description of the building they have in mind, specify every key dimension in plain text, press “Generate,” and immediately receive:
• A dimension-accurate floor-plan file
• A matching elevation model
• Downloadable outputs in at least DWG and IFC (plus a browser preview)
The engine must work solely from text-based prompts; no uploads or templates will be offered at this stage. Each result has to reach a highly detailed level suitable for permit drawings, with rooms, wall thicknesses, openings, materials and basic MEP placeholders already resolved.
The style selector inside the prompt must recognise Modern, Classical and Minimalist vocabulary and reflect it faithfully in the geometry and façade treatment. If a user asks for “a 15 m × 20 m minimalist concrete villa,” the service should respond with the correct footprint, proportional elevations and consistent material palettes. Also the plan pass should be editable by user so minor changes can be made.
Backend freedom is yours—stable diffusion–style LLM chains, parametric CAD scripting, or a hybrid pipeline—as long as response times stay under two minutes for a mid-size house and all geometry remains within a 2 mm tolerance of the stated dimensions. Host everything behind a clean React or Vue front end and include a REST endpoint so third parties can integrate the generator.
Acceptance criteria
• Upload a live URL plus Git repo for code review
• Supply a sample prompt library and the resulting DWG/IFC files for QA
• Pass an automated test set that checks dimensional accuracy and style compliance
I will sign off when the test set passes and the deployment guide lets me replicate the stack on my own AWS account without modification.
Location & Details
Apply on source →About this listing
This remote opportunity was imported from Freelancer and is shown here for discovery. To apply, follow the link to the original posting.