BISIDI AI image editor is a chat-based AI visualization platform that allows users to select part of an image, upload a real visual reference, and tell the AI what to replace or modify using natural language.
Generative AI can create impressive images in seconds. But editing an existing image still often begins with the same challenge: writing the right prompt.
BISIDI is taking a different approach.
Instead of learning prompt engineering, users can communicate with the image editor in a way that feels much closer to working with another person: say what you want, show an example, and point to where the change should happen.
The platform is designed for architects, interior designers, 3D visualization professionals, and e-commerce brands that need more control than traditional text-to-image generation provides.
What Is BISIDI?
BISIDI is an AI image editing and visualization platform focused on controlled changes to existing images.
Users can upload an architectural visualization, SketchUp or CAD view, interior image, product photograph or existing render and edit it through a chat interface.
If the change involves a specific object or material, the user can also upload a real reference image.
“Replace this chair with this one.”
“Use this stone on the facade.”
“Make the lighting warmer but keep everything else unchanged.”
Instead of describing every visual detail in a long prompt, the reference image provides the visual information while the text explains the action.
How Does the BISIDI AI Image Editor Work?
The workflow combines three types of input: natural language, visual references and image selection.
First, the user uploads the image they want to edit.
Then they select the exact object or area they want to change.
Finally, they can upload a reference image and describe the desired edit in normal conversational language.
Example 1: Replace Furniture
An interior designer wants to replace a chair in an existing room.
Rather than trying to describe the exact chair through a detailed prompt, the designer can:
- Upload the interior image.
- Select the existing chair.
- Upload a photograph of the real replacement chair.
- Type: “Replace this chair with this one.”
The goal is to change the selected object while keeping the surrounding room intact.
AI Image Editing With Real Reference Images
Visual references are particularly important when accuracy matters.
A person may be able to describe a general concept such as “light beige stone”, but that does not necessarily mean the AI will generate the exact material they have in mind.
With BISIDI, the material itself can become part of the instruction.
Example 2: Change a Facade Material
An architect can upload a building visualization, select part of the facade, and upload a photograph of a specific stone.
The instruction can be as simple as:
The same workflow can be applied to wood, flooring, fabrics, metals, wall finishes, and other architectural materials.
AI Rendering for Architects
For architects, generative AI becomes more useful when it can preserve the underlying project rather than constantly reinvent it.
BISIDI is designed to work with existing architectural imagery, including views from SketchUp, CAD, Revit and Archicad workflows.
Users can explore materials, lighting, landscaping, atmosphere, furniture, and seasonal conditions while keeping the geometry and proportions of the original project.
This moves AI rendering closer to the actual design process.
Instead of waiting for a completely new visualization every time a finish changes, architects can test alternatives directly on an existing image.
AI Interior Design With Real Furniture
The same principle applies to interior design.
Designers often work with real products from furniture manufacturers, online stores, Pinterest or product catalogues.
If the desired sofa, lamp or chair already exists as an image, describing it through text is unnecessary.
The user can simply upload the product photograph and show the AI where it should be placed.
Example 3: Place a Real Product in an Interior
Upload a room image, upload a photograph of the sofa, select the desired area and type:
“Place this sofa here and keep the rest of the room unchanged.”
This allows the conversation with AI to focus on the action rather than on writing a perfect description of the product.
AI Product Visualization for E-Commerce
BISIDI also applies this workflow to e-commerce and product visualization.
Furniture brands, lighting companies and online retailers traditionally rely on photography, CGI and 3D production to show products in different environments.
AI product visualization offers another workflow.
A brand can start with an image of a real product and create multiple environments around it.
The same chair could be visualized in a contemporary apartment, hotel lobby, restaurant or home office without organizing a separate photoshoot for every scene.
For e-commerce, the broader opportunity is interactive visualization.
A customer could eventually upload a photograph of their own room and see how a selected product might look inside that space.
How Is BISIDI Different From General AI Image Generators?
Most general AI image generators focus on creating a new image from a text description.
Professional image editing often requires the opposite.
Architects want to preserve geometry.
Interior designers want to use specific furniture.
Brands need the real product to remain recognizable.
BISIDI therefore focuses on controlled AI image editing with reference images, rather than simply generating a new scene from scratch.
“Can AI create a beautiful image?”
“Can AI change exactly what I want while keeping everything else intact?”
That is the problem BISIDI is built to solve.
As BISIDI founder and CEO Ievgeniia Leonova puts it:
“If the object, material or style already exists visually, why should someone spend time describing it? It is much more natural to show the AI what you mean, point to where you want the change and explain the action in one sentence.”
BISIDI is betting that the future of AI image editing will involve fewer complicated prompts and more natural communication: say it, show it, and point to it.
This article was published under HackerNoon’s Business Blogging program.
