In-Depth Review: Fenestra
What It Is
A browser-based AI rendering studio that takes sketches, CAD drawings as PDF, photos, or exported 3D models in FBX, GLB, and 3DM, and returns renders, illustrations, virtual staging, upscales to 8K, and short videos. It fronts 57 third-party image and video models under one subscription.
Who It's For
Architects and interior designers who want access to many current image models without managing several subscriptions or learning each tool's interface.
Why It Matters
The image models behind these tools change every few months, and a studio that buys one product is tied to whichever model it wrapped. Fenestra's argument is breadth: it publishes its model list, adds new ones as they ship, and lets you pick per job. A Realtime canvas also renders as you draw, which is closer to sketching than to prompting.
Problem Solved
Concept imagery arrives from several starting points, a pencil sketch one day and a Rhino export the next, and most tools accept only one. Fenestra takes sketches, PDFs of CAD drawings, photos, and 3D model files, and offers six modes across generation, editing, live canvas drawing, upscaling to 8K for client delivery, video, and 3D. Work stays in the browser with no GPU and nothing to install.
Audience Fit
Primary Users:
- Architect
- Interior Designer
- Designer
- Visualization Artist
Ideal Fit:
- Studios that want current image models without several subscriptions
- Concept work starting from hand sketches or PDF drawings as often as from models
- Teams needing an 8K image for client delivery from a rough starting point
- Anyone without a GPU workstation
Key Differentiator
One subscription across 57 named third-party models, plus a live drawing canvas
Workflow Fit
Project Phases Covered:
- Concept
- Schematic
Where It Plugs In:
Browser only. Models reach it as exported FBX, GLB, or 3DM files or as screenshots, since there is no plugin for any modeling application.
User Experience
Speed: Cloud generation with no install and no local GPU
Interface: Six modes covering image generation, editing, a Realtime canvas that renders as you draw, enhancement to 8K, video, and 3D models, with a model browser carrying preview thumbnails
Learning Curve: Gentle for single images. Choosing among 57 models is its own learning curve
Collaboration: Team plan covers up to five seats on a shared credit pool
Integrations: None. Revit, Rhino, and SketchUp are supported only through exported files or screenshots
Bonus Insight
The terms and conditions disagree with the pricing page in two places that affect client work. The pricing page says commercial use is included, while the terms grant a license for personal, non-commercial use only and bar commercial use except where the company specifically approves it. The pricing page also says there is no training on user content, a commitment the privacy policy does not repeat, and the terms take an unrestricted, irrevocable, perpetual, royalty-free license to user submissions, which is broad enough to cover training. The terms themselves are an unedited template, with a blank where the trial length should be, an empty address line, and a California residents clause inside a contract governed by the law of the United Arab Emirates. The company is Designtree AI LLC, registered in Sharjah, founded in late 2024 by Shaun McCallum. Uploads are processed by OpenAI, Black Forest Labs, Stability.ai, and Hugging Face, no data processing agreement is offered, and no retention period is stated. There are no independent reviews of this product anywhere, and its marketing line about proprietary models trained for architecture is contradicted by its own changelog, which lists third-party models only.
Need help implementing Fenestra?
Our advisory team has hands-on experience across 520+ AEC tools. Book a free AI strategy session to get specific recommendations for your firm.