RNDRS is an AI visualization tool for the architectural design field, primarily serving architects, interior designers, and landscape designers. Its core approach starts from images the user already has—a sketch, an exported model view, or an on-site photo—to generate concept-level renderings and progressively modify and compare options within the same project.
Purpose and Positioning RNDRS covers stages such as design development and facade options. It does not edit native CAD or BIM files, but instead processes images or screenshots exported from design software, so the model remains in the original software. It is suitable for conceptual exploration of scenarios such as exteriors, interiors, landscapes, aerial views, night scenes, and motion.
Main Capabilities
- Image input: Supports uploading JPEG, PNG, or WebP images up to 20 MB; you can also select a page from an unencrypted PDF up to 20 MB and 100 pages, and Studio will convert it into an image for conceptual visualization and editing.
- Local modifications: Use a brush, lasso, or polygon to select an area, then describe the modification requirements in text and check the overall result.
- Reference image guidance: Reference images can be attached to convey material, atmosphere, or style direction.
- Version management: Projects retain the source image and saved versions, allowing comparison of different directions and selection of earlier images to continue iterating.
- Review collaboration: Review links can be created for selected images and invited email addresses, with validity periods, collection of comments and decisions, and the ability to revoke review access.
- Output and presentation: Images can be downloaded, copies resized, or image packages exported, and results upscaled when necessary; still frames can also be turned into short video motion studies with cropping and mute controls.
How to Use Open RNDRS in a browser, upload a supported image format, describe the desired modifications, confirm the action cost before generating, and execute. No plug-ins or GPU are required. New accounts receive 5 credits for free, and a standard render consumes 1 credit; failed renders automatically refund credits. Paid plans start at $19 per month, or a one-time purchase of 300 credits for $99, and credits never expire. The monthly credit tiers shown on the page include 40, 100, and 300 credits. Generation uses credits, different operations and settings have different costs, and one credit does not necessarily correspond to one image.
Applicable Scenarios
- Interior: Starting from room photos or interior views, explore furniture, surfaces, and lighting, and check scale and circulation.
- Architecture: Use exported model views for material or daylight studies, explore facade overlays and atmosphere, while retaining the source image for comparison.
- Landscape: Use on-site photos or landscape views to explore paving, planting features, and seasonal atmosphere.
- District scale: Starting from aerialviews or master plans, explore conceptual spatial narratives.
Limitations and Notes
- AI may change geometry or introduce details, results need to be compared with the source image, and dimensions must be verified separately; generated images are not construction documents.
- Images and motion are both conceptual in nature, and geometry, details, and output dimensions must be checked before presentation.
- The upscaling feature can enlarge a selected image, but download dimensions and details need to be checked at the target presentation or print size.
- Image-to-video generates short clips, and camera movement is conceptual, not a measured walkthrough.
- Project pages require login, but anyone with the image delivery URL can access that image; cloud service providers process generation requests, and the privacy policy should be read before uploading confidential client materials.
- Paid plans allow commercial use under the terms, but users must have the necessary rights to the uploaded materials.
- Specific parameters such as supported image formats and upload limits are subject to the official page; information such as a Chinese interface and mobile support has not yet been provided.
