Grok Bot Tests Image Markup Feature to Circle Objects, Add Notes and Guide AI Photo Edits

Grok Bot is testing a new image-markup feature that could make it easier for users to communicate exactly what they want an AI assistant to change in a photograph. The feature allows users to circle specific objects, pin notes directly onto an image and paint over areas they want removed before sending the marked-up picture to their Bot.

The development was highlighted by technology observer Alex Michael, who shared a preview of the upcoming capability on X. If introduced widely, the feature could make image editing more precise by combining visual annotations with natural-language instructions. Instead of explaining an object’s location through lengthy text prompts, users would be able to mark the relevant area directly on the image.

How Grok Bot’s Image Markup Feature Works

The proposed feature focuses on making communication between users and AI assistants more visual.

Users will be able to upload or select a photograph and add different types of annotations to guide the Bot. These markings are intended to show the AI which parts of the image need attention and what kind of changes are required.

The reported capabilities include:

• Circle objects: Draw circles around specific objects or areas that need editing.
• Pin notes: Attach instructions or comments to particular locations in a photograph.
• Paint over objects: Mark areas that should be removed or changed.
• Send marked-up images: Share the annotated photograph with Grok Bot so it can understand the requested edits.

For example, a user could upload a photograph of a living room, circle an old sofa and add a note asking the Bot to replace it with a modern design. Another user could paint over unwanted objects and ask the assistant to remove them from the photograph.

The feature could also help with more complicated editing requests involving several different areas of one image.

Why Visual Annotations Could Improve AI Editing

AI image editing has traditionally relied heavily on written prompts. Users describe the object they want changed, explain its location and specify which parts of the original image should remain untouched.

However, text instructions can become complicated when photographs contain several similar objects or overlapping elements.

Consider a photograph containing multiple people, furniture items or products. Asking an AI model to change only one specific object can be difficult if the description does not clearly identify its location.

Image markup offers a more direct way to communicate those instructions.

By circling an object or placing a note next to it, users can provide additional visual context. The AI can then use the annotations alongside the written prompt to understand the intended edit.

This could reduce ambiguity and make editing workflows more efficient.

For professional users, the capability may be particularly useful when working with product photographs, marketing creatives, interior designs, presentations and social media images.

Grok Is Expanding Beyond Text-Based AI Assistance

The testing comes as xAI continues to expand Grok Bot’s capabilities beyond traditional conversational AI.

Grok Bot is designed to function as an AI agent that can handle tasks using connected services and its own computing environment. It can work with files, create documents, generate images and videos, interact with applications and carry out multi-step workflows.

Image markup would add another layer to this approach by making visual instructions easier to communicate.

The development also connects with xAI’s broader investment in its Imagine image-generation and editing ecosystem.

In August 2026, xAI introduced Imagine Image 2.0, highlighting more precise image generation and editing capabilities. The system includes tools such as a magic wand for editing selected regions, segmentation for selecting specific areas and background removal.

The company has also promoted natural-language image editing, allowing users to describe changes without manually creating masks.

Grok Bot’s proposed markup workflow could bring similar visual precision into conversations with AI agents. Rather than opening a separate editing interface and manually adjusting an image, users could potentially mark the photograph and delegate the requested changes directly to their Bot.

How It Could Help Designers and Content Creators

The feature could be useful across several creative and business workflows.

Product photography: Online sellers could circle a product component and request changes to its colour, background or appearance without describing every detail.

Social media content: Content creators could mark unwanted objects, highlight areas for correction or request changes to specific visual elements.

Interior design: Users could upload photographs of rooms and mark furniture, walls or decorative elements they want to replace.

Marketing and advertising: Marketing teams could annotate promotional images to communicate changes to backgrounds, product placement and other visual details.

Document and presentation design: Users could highlight specific elements in screenshots, diagrams and other visual material before asking Grok Bot to modify or explain them.

These are potential applications rather than confirmed workflows demonstrated in a public release. Their effectiveness will depend on how accurately the model interprets the markings and preserves the rest of the image.

Image Markup Is Becoming a Competitive AI Feature

Grok is not the only AI platform exploring ways to make image editing more interactive.

Adobe Photoshop has introduced AI markup tools that allow users to draw on images, add annotations and describe edits. These markings help guide the system towards specific regions and can be used to generate targeted variations.

Google has also been developing image-annotation capabilities for Gemini, allowing users to circle objects, draw arrows and add notes to explain which parts of an image they want the AI to examine or modify.

The broader direction is clear: AI image tools are moving beyond simple text prompts towards interfaces that combine drawing, selection, annotation and natural-language instructions.

This approach gives users more control over the editing process while reducing the need for advanced design skills.

For xAI, bringing these capabilities into Grok Bot could make image editing part of a wider AI-agent workflow rather than a standalone creative tool.

Feature Is Still Being Tested

Despite the early preview, Grok Bot’s image-markup capability has not been confirmed as a generally available feature.

The information currently available comes from a report shared by Alex Michael on X. The preview describes the ability to circle areas, pin notes and paint over parts of photographs before sending them to a Bot.

There is no confirmed public release date, supported-device list or detailed explanation of how the feature will work across different image formats.

It is also unclear whether the final version will support multiple annotations, different brush sizes, undo controls or more advanced selection tools.

The feature could change before a wider rollout, and its availability should not be confused with Grok’s existing image-generation and editing capabilities.

The Bigger Picture

The image-markup test reflects a broader shift in how people interact with AI systems. Instead of relying entirely on written instructions, users are increasingly being given visual tools to communicate their intentions.

This is particularly important for image editing, where the location of an object can be just as important as the requested change. Combining annotations with natural-language prompts could make AI tools easier to use for people without professional editing experience.

For xAI, the feature could strengthen Grok Bot’s position as a multipurpose assistant capable of handling creative tasks alongside research, productivity and software interactions. It also places the company in a growing competition with Google and Adobe to make AI-powered visual editing more accessible.

Looking Ahead

The next important development will be whether xAI moves the image-markup feature from testing into a public release. A successful rollout could give Grok Bot users a more direct way to request image changes, particularly when working with complicated photographs that require precise edits. Details about availability, editing controls and the accuracy of the final results will determine how useful the feature becomes.

More broadly, visual annotations could become a standard part of AI assistants as companies look for better ways to connect human instructions with machine understanding. If Grok can combine accurate object selection, contextual notes and reliable image editing, it could make creative work more accessible to everyday users and reduce dependence on traditional editing software for simple tasks.