Veras AI Rendering for Revit, SketchUp, and Rhino: Complete Guide

Most AI image tools operate outside your design environment. You write a text prompt in a browser, wait for a result, and then try to reconcile that image with the 3D model sitting in your CAD application. Veras by EvolveLab takes a different approach. It runs as a plugin directly inside Revit, SketchUp, and Rhino, capturing your current viewport and using AI to transform it into a styled render while preserving the geometry you already modeled.

This guide covers everything from installation and first render to advanced style control, batch processing, and real project workflow integration.

What Veras Is and How It Works

Veras is an AI-powered rendering plugin developed by EvolveLab, a company known for building productivity tools for AEC professionals. Unlike standalone AI image generators such as Midjourney or Stable Diffusion, Veras works from within your modeling software. It takes a screenshot of your current 3D view, sends it to cloud-based AI models along with your style instructions, and returns a rendered image that follows the geometry visible in your viewport.

The core difference from text-to-image tools is that Veras uses your model as a structural guide. The AI does not invent building forms or guess at spatial relationships. It works with the massing, openings, roof lines, and context already present in your scene. You control how closely the output adheres to your geometry through a strength slider, and you direct the visual style through prompts and presets. This places Veras in a unique space between traditional rendering engines and pure AI generators - creative flexibility with your actual design as the foundation.

Installation and Setup for Each Platform

Getting Veras running requires an EvolveLab account and a subscription or trial license. The installation process varies slightly depending on your host application.

Revit

1. Download the Veras installer for Revit from the EvolveLab website
2. Close Revit, run the .exe installer (it detects your installed Revit versions automatically)
3. Launch Revit and look for the Veras tab in the ribbon
4. Sign in with your EvolveLab account when prompted

Veras supports Revit 2022 and later. Use a 3D perspective or camera view rather than a plan or section, as the AI needs spatial depth information to produce convincing results.

SketchUp

1. Download the extension from the EvolveLab website or the SketchUp Extension Warehouse
2. In SketchUp, go to Extensions > Extension Manager > Install Extension, select the .rbz file
3. Restart SketchUp and access Veras from the Extensions menu or toolbar

SketchUp users benefit from Veras especially during early design phases when the model is still simple massing. The AI fills in material detail and environmental context that would otherwise require extensive setup.

Rhino

1. Download the Rhino plugin from the EvolveLab website
2. Use the PackageManager command or manually place plugin files in your Rhino plugins folder
3. Restart Rhino and type “Veras” in the command line to launch the panel

Rhino integration works with both Rhino 7 and Rhino 8. Grasshopper users can incorporate Veras into parametric workflows, though the rendering happens on the current viewport capture rather than on individual Grasshopper geometry states.

The Rendering Workflow Step by Step

The rendering process follows a consistent pattern across all three platforms.

Step 1: Set Up Your View. Navigate to a 3D perspective view with clear spatial depth. Avoid extreme close-ups - the AI performs better when it can interpret the scene as a complete architectural composition.

Step 2: Open the Veras Panel. Launch Veras from the plugin tab or menu. The panel shows your current viewport as a thumbnail preview.

Step 3: Write Your Prompt. Describe the visual style you want - materials, lighting, atmosphere, and context. For example: “Modern residential exterior, warm evening light, timber cladding with dark metal accents, lush garden landscaping, overcast sky.”

Step 4: Choose a Style Preset (Optional). Use a built-in preset alone, combine it with your text prompt, or skip presets entirely.

Step 5: Adjust Geometry Override Strength. This slider controls how closely the AI follows your model geometry. Higher values stick closer to your model; lower values give the AI more creative freedom.

Step 6: Click Render. Veras captures your viewport and sends it to the cloud. Render times range from 15 to 60 seconds depending on server load and resolution.

Step 7: Review and Iterate. If the composition needs adjustment, rotate your view and render again. If the style needs refinement, adjust your prompt or strength settings. Each render consumes one credit.

Style Presets and Custom Style Control

Veras ships with style presets organized by visual category - a fast way to get started before mastering prompt writing.

Common preset categories include:

• Photorealistic - Clean, natural lighting with realistic materials
• Sketch and Watercolor - Hand-drawn looks that work well for early design presentations
• Cinematic - Dramatic lighting and atmosphere, useful for competition imagery
• Minimalist - Clean, bright, Scandinavian-influenced aesthetics
• Night and Moody - Evening scenes with artificial lighting and atmospheric effects
• Landscape and Context - Emphasizes site context, vegetation, and environmental integration

Beyond presets, the real power of Veras is in custom prompts. Effective prompts should address four dimensions:

1. Materials and finishes - “Exposed concrete walls, floor-to-ceiling glass, brushed brass hardware”
2. Lighting and time of day - “Late afternoon golden hour, soft shadows, warm ambient light”
3. Atmosphere and mood - “Serene, minimal, gallery-like atmosphere”
4. Context and landscaping - “Mature oak trees, gravel courtyard, low stone walls”

You can also reference specific architectural photographers in your prompts. Phrases like “in the style of Iwan Baan photography” or “Mir visualization aesthetic” can meaningfully shift the visual quality.

Understanding Geometry Override Strength

The geometry override strength slider is what separates Veras from pure text-to-image tools, and understanding it properly is essential for getting useful results.

High strength (0.7 to 1.0): The AI follows your model geometry closely. Window positions, wall lines, and massing remain accurate. The AI adds material detail and environmental context but does not alter the building form. Use high strength for client presentations and design reviews.

Medium strength (0.4 to 0.6): The AI respects the general massing but may soften edges, add vegetation, or reinterpret material boundaries. This range works well for concept exploration where you want suggestions while staying grounded in your geometry.

Low strength (0.1 to 0.3): The AI uses your viewport mainly as a compositional guide and may significantly alter forms or add elements not in your model. Useful for very early ideation with simple box models.

Start at 0.6 to 0.7 for your first render of any view, then adjust based on whether you need more accuracy or more creative interpretation.

Comparison with Traditional Rendering Engines

Architects already using V-Ray, Enscape, Lumion, or Twinmotion naturally want to know where Veras fits in the toolkit. The short answer is that it does not replace these tools - it serves a different purpose.

V-Ray and traditional path-tracing renderers produce physically accurate lighting, reflections, and material behavior. They require detailed material setup, lighting rigs, and significant render time. Veras cannot match this level of accuracy.

Enscape and Twinmotion offer real-time rendering with fast setup and good-quality output. They require material assignment but far less than V-Ray. For walkthroughs and polished client visuals, they remain excellent choices.

Veras fills the gap where you need a styled visual but do not have time to set up materials, lighting, and environment in a traditional renderer. Use Veras for speed and creative exploration in early and mid-design phases. Switch to Enscape, Twinmotion, or V-Ray when you need physical correctness for final deliverables.

Pricing and Plans

Veras operates on a credit-based subscription model. Each render consumes credits, and plans differ in how many credits you receive per month.

As of early 2026, EvolveLab offers several tiers:

• Free Trial - A limited number of credits to test the tool before committing
• Individual Plans - Monthly or annual subscriptions with a set number of renders per month, suitable for solo practitioners
• Team Plans - Higher credit limits with multi-seat licensing for firms
• Enterprise - Custom pricing for large organizations with volume needs

Pricing changes periodically, so check the EvolveLab pricing page for current rates. The credit cost per render depends on the output resolution - higher resolutions consume more credits per image.

Because each render costs a credit, refine your view and prompt before hitting render. Developing good prompting habits saves money over time.

When to Use Veras vs Other AI Tools

Knowing when each tool makes sense helps you pick the right one for the task at hand.

Use Veras when:

• You have a 3D model and want the render to follow your actual geometry
• You need quick concept visuals without leaving your modeling software
• You want to show clients multiple style directions for the same building form
• You are producing competition boards and need atmospheric imagery fast

Use Midjourney or DALL-E when:

• You are in pre-design and do not have a 3D model yet
• You want reference imagery or mood boards from text descriptions alone
• You are exploring abstract spatial concepts

Use Stable Diffusion (with ControlNet) when:

• You want free, unlimited local generation with geometric control
• You need fine control over the AI pipeline (custom models, LoRA weights)

Use traditional renderers (V-Ray, Enscape, Lumion) when:

• You need physically accurate material representation
• Clients expect photorealistic accuracy of the final design
• You need consistent output across dozens of views

Batch Rendering for Multiple Views

One of Veras’s most practical features for real project work is the ability to render multiple views in sequence. Instead of manually switching between views and clicking render each time, you can queue up several camera positions and let Veras process them as a batch.

The workflow for batch rendering:

1. Set up named views or cameras in your modeling software (Revit camera views, SketchUp scenes, or Rhino named views)
2. In the Veras panel, select the views to include in the batch
3. Apply the same prompt and style settings across all views, or customize per view
4. Start the batch process - Veras renders each view sequentially

This is particularly useful for design review packages where you need exterior, interior, and context shots all in a consistent style. Instead of spending hours on rendering setup for an early-stage review, you can produce a cohesive set of views in minutes. Keep in mind that each view consumes one credit, so plan batches thoughtfully on limited tiers.

Real Project Workflow Examples

Understanding where Veras fits in actual project timelines makes it more useful than treating it as a standalone tool. Here are three common workflow scenarios.

Scenario 1: Competition Entry

You have a massing model in Rhino and two weeks to produce a competition board. Instead of spending the first week on material setup, you use Veras to generate atmospheric concept imagery from your massing model during the first few days. The design team evaluates spatial compositions early, and final hero images can be refined with traditional rendering later.

Scenario 2: Client Design Review at Schematic Design

Your Revit model has basic wall types and generic materials, and the client meeting is in two days. Rather than rushing material assignments in Enscape, you run Veras on your key perspective views with a prompt describing the intended material palette. The client sees styled visuals that communicate design intent without a full day spent on rendering setup.

Scenario 3: Social Media and Marketing Content

Your firm wants to post project visuals regularly. Running existing 3D models through Veras with different seasonal prompts (summer garden, winter evening, spring morning) generates a library of varied imagery from the same base geometry, far more efficiently than re-rendering from traditional engines for each variation.

Limitations You Should Know

Veras is a powerful tool, but it has clear boundaries that you need to understand to use it effectively and set appropriate expectations with clients and colleagues.

Geometric accuracy is approximate, not exact. Even at high geometry strength, the AI may shift window mullion patterns, alter facade rhythms, or slightly modify proportions. If a client is reviewing specific design details, Veras output may misrepresent them.

Interior rendering is less reliable. Interiors with complex furniture layouts and tight spaces challenge the AI more than exterior views with clear massing and sky context. Interior results often require more iteration.

Consistency across views is not guaranteed. Each Veras render is an independent AI generation. The same prompt may produce slightly different material interpretations on different views, which can be noticeable in formal presentation sets.

Cloud dependency. Every render requires an internet connection and EvolveLab’s servers. Offline work, server outages, or high demand can all prevent or delay rendering.

Credit cost adds up. For firms rendering dozens of views weekly, subscription costs can become significant. Track your usage during the trial to estimate realistic monthly costs.

Best Practices for Getting Great Results

After working through the basics, these practical tips will help you get consistently better output from Veras.

Compose your model view before rendering. A well-composed perspective with clear foreground, middle ground, and background produces dramatically better results than a flat elevation view. Think of your viewport as a photograph you are handing to the AI for enhancement.

Be specific in your prompts. Vague prompts like “nice modern house” produce generic results. Detailed prompts like “two-story residential, charred timber cladding, zinc standing seam roof, concrete retaining walls, native grass landscaping, soft overcast daylight” give the AI enough direction to produce something distinctive.

Use the same prompt across a batch for consistency. When rendering multiple views of the same project, use identical prompts to maintain a cohesive visual language across the set.

Start with medium geometry strength and adjust. Beginning at 0.6 gives you a good balance. Move up if the AI takes too many liberties with your geometry, or move down for more stylistic freedom.

Save your successful prompts. When you find a prompt that works, save it in a text file or your firm’s project template. Prompt libraries save time across projects.

Render at the appropriate resolution. Use lower resolution for quick iterations. Save high-resolution renders for final selections. This conserves credits and speeds up the iteration cycle.

Combine with post-processing. Veras output benefits from light post-processing in Photoshop. Adjusting levels, adding a vignette, or compositing text on top of the render turns a good AI image into a polished presentation graphic.

Getting Started with AI Rendering in Your Practice

Veras represents a practical entry point into AI-assisted visualization that works within the tools architects already use daily. Unlike browser-based AI generators, it respects your geometry, integrates into your modeling workflow, and produces results grounded in your actual design.

The best way to evaluate it is to install the trial, pick a current project, and run a few renders with different prompts and geometry strength settings. Within 30 minutes you will know whether the tool fits your workflow.

If you want to build a stronger foundation in the software tools behind modern architectural practice, explore our course catalog at Archgyan Academy for structured learning paths in Revit, SketchUp, and computational design.