Good morning. Grok Imagine just dropped Agent Mode. One sentence and it plans, generates, edits, and stitches an entire video project for you. And a viral prompt is turning old damaged photos into DSLR-quality portraits using ChatGPT. Have you tried Agent Mode yet? Hit reply and tell me.
If you're new here, this is where we cover what's actually working in AI video before everyone else finds out. The full breakdown and tutorial is below.
🎬 Grok Imagine Agent Mode Just Dropped
xAI launched Grok Imagine Agent Mode in beta this week. Instead of generating one image or video at a time, you describe a full creative project and the agent handles everything. "Make a 1-minute short film." "Generate a UGC product shoot across 5 SKUs." "Build a complete manga set." It plans, generates, edits, converts to video, and stitches clips together on its own.
The whole thing runs on an infinite canvas at grok imagine on desktop. It comes with preset workflow templates for short films, product stories, brand identities, and world-building. The agent keeps context across every step so edits build on what came before instead of starting fresh. You need a paid Grok account to access it. Toggle it on at the bottom left of the input field.
Old Photos Are Becoming 4K Portraits With One Prompt
Created with GPT Images 2.0
A prompt is going viral right now that turns damaged, faded old photos into sharp DSLR-quality portraits using ChatGPT. Upload your photo and paste this: "Restore this old photo into a professional portrait of DSLR quality colour and detail, using an advanced upscaling algorithm comparable to results from a Canon EOS R6 II. Ensure the restored image looks natural, retains exact facial features, has great clarity."
It works because ChatGPT Images 2.0 can now handle text and detail rendering at a level previous models couldn't. The key is telling it to retain exact facial features. Without that, the AI tends to "improve" people into slightly different people. This works on black-and-white photos, cracked prints, faded snapshots. If you've got a shoebox of old family photos, this is worth 5 minutes of your time.
How to Use Grok Imagine Agent Mode
Created with Grok Imagine Agent
In this example we will create an AI product video using Grok Imagine. Go to grok imagine click on Agent on the bottom panel, upload an image of the product (like a tshirt), and describe the avatar.
Attach a product image and use this prompt to follow along: In 9x16 format of streetwear vibe woman model with long hair, sharp features, natural realistic skin with visible pores, wearing this exact shirt
Grok will create an avatar wearing your product. Once it’s generated hit “Make video” in the bottom right. Optionally you can add specific instructions on what you want to happen in the video.
Takeaway: This is the closest thing to a full AI video production system inside a single interface. Instead of bouncing between ChatGPT for scripts, Midjourney for images, and Kling for video, Agent Mode handles the whole pipeline in one workspace.
🧪 Tool of the Day:
Grok Imagine Agent Mode: Describe an entire creative project and let an AI agent plan, generate, edit, and stitch it together on an infinite canvas. Beta, paid accounts only.