Midjourney acquires astrology app Co-Star to expand beyond image and video generation.

What Happened

Midjourney, the AI research lab renowned for its state-of-the-art text-to-image models, has acquired the popular astrology application Co-Star. This marks a highly unusual but strategic pivot for Midjourney, signaling an expansion beyond foundational media generation utilities into the consumer application ecosystem.

Technical Details

While Co-Star presents as a mystical consumer app, under the hood, it operates as a sophisticated data and natural language generation (NLG) engine. It maps NASA planetary data against user-provided birth metrics to generate highly structured, hyper-personalized daily text updates. By acquiring Co-Star, Midjourney gains access to a massive, proprietary dataset of user-specific behavioral data, psychological traits (as categorized by the app's ontology), and daily engagement metrics. This structured, high-signal data is exceptionally valuable for fine-tuning multimodal systems for extreme personalization, moving beyond zero-shot prompting into continuous context.

Why It Matters

From an engineering and product perspective, this is a strategic acquisition of both a data moat and an active distribution channel. Midjourney currently operates primarily as a pull-based utility (users input prompts via Discord or web). Co-Star provides a sticky, push-based mobile interface with millions of daily active users (DAUs). Combining Midjourney's generative prowess with Co-Star's algorithmic personalization pipeline paves the way for dynamically generated, multimodal content. We are likely looking at the foundation for context-aware generative media—where high-fidelity images or videos are generated proactively based on a user's psychological profile, daily context, and historical engagement, rather than requiring an explicit user prompt.

What to Watch Next

Monitor how Midjourney integrates its image and upcoming video generation models into the Co-Star mobile architecture. The immediate technical hurdle will be optimizing inference latency and compute costs to generate high-fidelity visual assets on the fly for a massive DAU base. Additionally, watch for signs of Midjourney developing its own conversational agent, utilizing Co-Star's proprietary data to train an empathetic, personalized AI companion.