xAI Debuts Imagine Agent in Grok with Open Canvas AI Workspace

xAI has begun rolling out Imagine Agent inside Grok Imagine on the web, transforming the tool from a single-shot image generator into a fully autonomous creative workspace. The agent operates within an open canvas, accepting broad, multi-step requests such as “generate a 1-minute film” or “generate a complete manga set.” It reasons through the creative brief, drafts a scenario, generates scene clips, stitches them into a sequence, and even produces a companion poster image. For product shoots, it can generate assets across multiple SKUs and fuse images into composite scenes. The feature is available in beta to Grok Heavy and Super Grok subscribers who already have Grok Imagine privileges, with a wider rollout expected soon.

This launch positions xAI against OpenAI’s Images 2.0, Meta’s Vibes platform, and Google’s AI Studio, all racing to ship agentic creative surfaces. By layering an autonomous agent on the canvas, xAI extends the multi-agent architecture introduced in Grok 4.20 into a consumer-facing product. Users can now delegate entire creative projects—campaigns, films, marketing assets—without leaving Grok, shifting from prompt-by-prompt iteration to hands-off delegation. An official announcement from xAI is likely imminent.