Sherpa is an AI rendering tool for architects and interior designers that works from an existing work render or SketchUp export rather than a text prompt. Uploading a base image returns a presentation-ready visual without any other render engine in the pipeline.
Instead of prompt writing, output is driven by a control layer: luminosity, time of day, colour temperature, artificial lighting, presence of people, floor and wall materials, and interior styling. The same camera and the same export can be re-run with a single control changed, and every version stays grouped with its original.
A finished image can be worked on further: adjust one element while keeping the rest intact, pull four fixed detail shots (material, furniture, light, styling) from one render, or upscale to 4K after the composition is settled.
A free Sherpa Render extension is available on Trimble’s SketchUp Extension Warehouse, rendering straight from the viewport without an export step. Input and output images are not used for model training, remain the property of the user, and are deleted after 30 days.
Sherpa is designed to be used during the concept and pitch stage. It’s developed by Amsterdam-based AI-brouwerij (AI Brewery), runs in the browser and accepts exports from SketchUp, Revit and ArchiCAD. New accounts get five free renders without a credit card. Rental options start at €49/month.
More on Sherpa’s website.
The main issue with AI rendering tools right now is cost. They’re essentially all subscription and/or credit based, and serious use burns through credits fast.
If AI rendering is supposed to replace traditional rendering in the long term, this simply doesn’t scale. Local AI seems key, but the current options — ComfyUI, ControlNet, custom models, LoRAs, nodes and complex workflows — are still far too technical and fragmented for everyday production.
What’s missing is the simplicity of current AI render tools, but running locally with virtually unlimited iterations.