Maket AI for Architects: Automated Residential Floor Plan Generation

Residential floor plan design is a balancing act between client preferences, building codes, site constraints, and livability. Architects spend days iterating on plan layouts before settling on two or three options to present. Maket AI compresses that iteration cycle by generating constraint-based floor plan options in minutes, giving you a wider pool of starting points before you invest hours in manual refinement.

This guide walks through what Maket AI actually does, how to set up a residential project, the constraint system that drives generation, style controls, export workflows, pricing, limitations, and when this tool genuinely helps versus when manual sketching is faster.

What Maket AI Does

Maket AI is a generative design platform focused on architectural floor plan creation. You describe what you need - the number of rooms, their sizes, adjacency relationships, and site boundaries - and the AI generates multiple floor plan options that satisfy those constraints. The platform runs entirely in the browser with no local software installation.

The core workflow follows a straightforward pattern. You create a project, define the building footprint, input your room program with area requirements, set relationships between spaces, choose a style direction, and generate. Within minutes, the system returns floor plan options you can browse, compare, and refine. Each plan includes room labels, approximate dimensions, door and window placements, and area calculations.

What makes Maket AI distinct from a generic image generator is that every output is a spatially valid floor plan. Rooms do not overlap, circulation paths connect spaces logically, and the total area stays within your defined footprint. It is solving a spatial allocation problem, not generating a picture that looks like a floor plan. The output is meant to be a working starting point you can import into your CAD or BIM tool.

The platform also includes a natural language interface. You can describe a project in plain text - “a 2000 square foot single story home with 3 bedrooms, 2 bathrooms, open concept kitchen and living room, attached 2-car garage” - and Maket AI interprets that into constraints and generates plans accordingly. This conversational approach lowers the barrier for early-stage exploration when you do not yet have precise programmatic requirements.

How Constraint-Based Generation Works

Maket AI uses a constraint solver at its core, with a user interface built around natural language and visual editing rather than technical parameter lists.

When you define a project, the system builds an internal model with several layers of constraints:

• Boundary constraints define the outer envelope - your lot line, building footprint, or the perimeter of an existing structure
• Programmatic constraints specify rooms you need with type, target area range, and quantity
• Adjacency constraints define spatial relationships - kitchen connects to dining, master bath is accessible from master bedroom
• Separation constraints keep spaces apart - bedrooms separated from utility areas, guest rooms isolated from the master suite
• Orientation preferences specify which rooms face which direction for solar gain, views, or privacy
• Fixed elements mark non-negotiable positions like existing stairwells, column grids, or required entry points

The generator explores the solution space defined by these constraints, producing 5 to 20 plan options depending on complexity. Loosely constrained projects generate more options. Tightly constrained projects produce fewer but more targeted results. Each plan is scored on area utilization, adjacency satisfaction, circulation efficiency, and orientation alignment.

Setting Up a Residential Project Step by Step

Here is a practical walkthrough for a 2,400 square foot, two-story home.

Step 1: Define the Site and Footprint

Create a new project, select residential as the type, and define the building footprint. You can draw it with the polygon tool or describe it: “rectangular footprint, 40 feet wide by 30 feet deep, long axis running east-west.” For two-story homes, define each floor as a separate plan level.

Step 2: Input the Room Program

Define each room with its type, target area, and quantity:

Ground Floor: Entry foyer (60-80 sq ft), living room (300-350 sq ft), kitchen (180-220 sq ft), dining area (150-180 sq ft), powder room (30-40 sq ft), mudroom (50-70 sq ft), attached 2-car garage (400-450 sq ft).

Second Floor: Master bedroom (250-300 sq ft), master bathroom (80-100 sq ft), walk-in closet (50-70 sq ft), bedroom 2 (150-180 sq ft), bedroom 3 (130-160 sq ft), full bathroom (60-80 sq ft), laundry room (40-60 sq ft).

You can input these individually or use natural language: “Ground floor with open concept kitchen-dining-living of about 650 square feet combined, foyer, powder room, mudroom, and two-car garage. Second floor needs a master suite with walk-in closet and ensuite, two additional bedrooms sharing a full bath, and a laundry room.”

Step 3: Set Adjacency Relationships

Define the relationships that matter to the design:

• Kitchen must connect to dining area (hard constraint)
• Dining must connect to living room (hard constraint)
• Mudroom must connect to garage and kitchen (hard constraint)
• Master bedroom must connect to bathroom and walk-in closet (hard constraint)
• Bedrooms 2 and 3 should be near shared bathroom (soft constraint)
• Laundry should be near bedrooms (soft constraint)
• Garage should not connect directly to living areas (separation constraint)

Hard constraints eliminate any plan that violates them. Soft constraints influence scoring without disqualifying plans. Use hard constraints sparingly for non-negotiable relationships.

Step 4: Set Orientation and Generate

Specify compass orientation for the front of the house, then set room preferences. Living room and kitchen on the south side for afternoon light. Bedrooms on the east for morning light. Garage on the north where it blocks cold winds without sacrificing window area.

Generate your first batch and review: Does the circulation make sense? Are room proportions reasonable (a 300 sq ft living room should not be 30x10)? Do adjacencies flow naturally? Where did the stairs land? If results are not viable, adjust constraints and regenerate.

Style and Aesthetic Controls

Style controls in Maket AI affect the generation logic itself, not just cosmetics applied afterward.

Architectural style presets include Modern, Traditional, Craftsman, Mediterranean, and Contemporary. Selecting Modern biases the generator toward open floor plans, larger windows, and simpler geometries. Traditional presets produce more defined rooms with clear separation between spaces.

Open vs. closed plan preference controls room separation. At the open end, kitchen, dining, and living merge into a single zone. At the closed end, each space gets its own walls and doorways.

Circulation style determines whether the generator favors hallway-based circulation (rooms accessed from a central corridor) or flow-through circulation (rooms connect directly to each other). Setting the same room program with different style presets produces genuinely different plan organizations, not just different labels on the same layout.

Export Formats and CAD Integration

Maket AI supports several export paths for moving plans into production tools:

DXF export creates a 2D CAD file compatible with AutoCAD, BricsCAD, or any DXF application. The export includes walls as line work, room labels, door swings, and basic dimensions. This is the most universally compatible option.

PDF export generates presentation-quality floor plans suitable for early client meetings. Image export (PNG/JPG) produces raster files for quick sharing.

Revit integration works through the DXF pathway. Export your selected plan as DXF, import it into Revit as a linked CAD file, and use it as an underlay for modeling walls and placing doors. The DXF provides accurate wall positions and room boundaries to trace over.

The export quality is adequate for schematic design. Walls export as simple line geometry, not parametric wall objects. Door placements show swings but carry no specification data. You are getting an accurate spatial layout that requires development in your BIM tool.

Pricing and Plan Tiers

Maket AI operates on a subscription model. The free tier lets you create limited projects and generation counts - enough to evaluate the tool on one or two test cases. The Pro tier (around $29-49/month) removes limits, enables full exports, and unlocks advanced constraints. Team plans add collaboration and per-seat licensing.

Check the Maket AI pricing page for current rates, as pricing changes frequently for AI tools. For a solo practitioner doing 8-12 residential projects per year, the Pro tier pays for itself if it saves even a few hours per project on schematic exploration.

Maket AI vs. Finch 3D

Both tools generate floor plans from constraints, but they serve different scales and workflows.

Maket AI strengths: Natural language input with minimal training needed. Style presets that produce architecturally varied results. Optimized for residential design with room types and adjacencies that match how residential architects think. Fast setup - blank project to generated options in 15-20 minutes.

Finch 3D strengths: Deeper BIM integration, particularly with Revit. More granular control over structural grids and multi-story coordination. Better suited for larger multi-family residential, commercial, and institutional projects.

Choose Maket AI for single-family homes, small multi-family (2-4 units), or early feasibility studies where you want many layout options quickly. Choose Finch 3D for mid-to-large residential (12+ units), mixed-use buildings, or projects where structural grid alignment drives the plan.

When to Use AI vs. Manual Design

Maket AI adds value when you need 10+ layout variations for a feasibility study, when clients keep changing the room program, when optimizing a spec home for a standard lot, or when site geometry makes layout non-obvious. It also helps early-career architects build spatial intuition by studying many generated configurations.

Manual design is better when a strong design concept drives the plan (a central courtyard, spiral circulation), the site has complex topography requiring section-driven design, the program is unusual, or you already know the configuration you want. Maket AI excels at breadth - exploring many configurations of a well-defined problem. Manual design excels at depth - developing ideas that require creative judgment no constraint solver can replicate.

Renovation and Addition Projects

Maket AI handles renovation projects with additional setup. For additions, define the existing building as a fixed element, mark the addition zone as flexible, and set connection points between old and new. For gut renovations, use the existing footprint as your boundary and structural elements as fixed, then generate new interior layouts within the remaining space.

Partial renovations require defining which rooms stay and which can change. Draw the renovation zone as a sub-boundary with fixed adjacent rooms as connection constraints.

The key limitation with renovation work is that Maket AI does not understand construction constraints like plumbing stack locations, header heights, or mechanical routing. Moving the bathroom 20 feet from the existing stack adds significant cost, but the AI does not know that. You need to encode this knowledge as fixed elements or adjacency constraints before generating.

Limitations to Know

Room proportions are not always livable. The generator satisfies area requirements but sometimes produces awkward aspect ratios - an 18x8 bedroom meets the square footage target but is not a usable room.

Structural reality is abstracted. A generated plan might show a 25-foot open span that would require significant engineering. The tool assumes walls are thin partitions unless told otherwise.

Mechanical systems are invisible. Plumbing, HVAC, and electrical routing are not considered. Plans might place bathrooms at opposite ends of the house on different floors - spatially valid but mechanically expensive.

Building codes are your responsibility. Maket AI does not automatically enforce egress widths, window sizes for emergency escape, accessibility requirements, or ceiling heights. Encode these as constraints or verify compliance after generation.

No roof or section generation. The platform works in plan only. If the roof geometry drives interior spatial experience - vaulted ceilings, dormers, loft spaces - design those in section and use Maket AI only for plan organization.

Best Practices for Getting Good Results

Define room types precisely. Use “master bedroom,” “child bedroom,” and “guest bedroom” with different area ranges and adjacency requirements instead of generic “bedroom.” This forces the generator to differentiate between spaces.

Set circulation as a first-class constraint. Define entry connections, stair placement, and the main circulation spine. Leaving circulation to the generator often produces technically valid but experientially strange paths through the house.

Generate in batches with different priorities. Run one batch optimizing for compactness, another for natural light, and a third with relaxed area constraints. Comparing across batches reveals trade-offs a single run does not expose.

Use natural language for exploration, then manual constraints for refinement. Describe the project in plain text for a quick baseline, then set precise constraints for targeted second-round generation.

Always verify room proportions before exporting. Scan every generated plan for rooms with aspect ratios worse than 1:2. A 150 sq ft bedroom at 10x15 is fine. At 6x25, it is unusable.

Save constraint templates. If you design custom homes regularly, build a library of constraint sets for common programs - three-bed ranch, four-bed colonial, open-concept modern. Each template saves 30-60 minutes of setup on the next similar project.

Getting Started

The most practical way to evaluate Maket AI is to use it on a real project in early schematic design. Take your actual room program, set up the constraints, and compare generated options to what you would sketch by hand. Start with the free tier to test the workflow. If the generated plans surface layout ideas you would not have considered, or if the iteration speed saves meaningful time during client meetings, move to the paid tier.

For architects building a broader toolkit of AI and computational design skills, explore our course catalog at Archgyan Academy. We cover Revit, BIM workflows, and the growing landscape of AI tools in architecture with practical, hands-on instruction designed for working professionals.

The floor plan remains the most consequential drawing in residential architecture. It determines how a family moves through their home, how light enters rooms, and how private spaces feel separate from public ones. Maket AI does not make those decisions for you. It shows you more possibilities, faster, so the decision you make is informed by a wider exploration than manual iteration alone could achieve.