• AiNews.com
  • Posts
  • MidJourney Introduces Moodboards and Custom AI Image Models

MidJourney Introduces Moodboards and Custom AI Image Models

A vibrant artist's workspace with a large digital screen displaying a colorful moodboard of inspiring images. A person sits at a desk using a tablet to refine AI-generated art, while another person collaborates by pointing at the screen. The modern studio is well-lit, with sketchbooks, laptops, and art tools organized in the space, emphasizing creativity and teamwork in a tech-driven environment.

Image Source: ChatGPT-4o

MidJourney Introduces Moodboards and Custom AI Image Models

MidJourney, the widely used AI image generator boasting over 19 million users, has unveiled new tools aimed at enhancing creativity and personalization. These updates include Pinterest-inspired moodboards and support for multiple personalization profiles, enabling users to tailor the AI’s image generation capabilities to their unique artistic visions.

Moodboards: Inspiring Creativity

The new moodboards feature allows users to upload curated collections of images to inspire the creation of new art. By analyzing the diversity and complexity of the uploaded images, MidJourney's AI develops a unique style profile, remixing visual elements into a cohesive aesthetic.

This feature is designed to help users streamline their creative process, providing a foundation for generating art that aligns closely with their intended style.

Personalization Profiles: Tailoring Creativity

In addition to moodboards, MidJourney now supports multiple personalization profiles. Users can create and switch between various customized versions of the platform’s latest AI model, version 6.1. This functionality is ideal for managing different projects, personal styles, or collaborative efforts.

Key features of personalization profiles include:

  • Naming and organizing profiles.

  • Setting one or more profiles as defaults.

  • Tracking images tied to specific profiles.

These features make it easier for users to manage their workflows, especially when working on multiple projects or within teams.

Streamlined Custom Model Training

MidJourney has simplified the process of training custom models by improving the image ranking system. Previously, users needed to rate 200 images to personalize their model, but now only 40 ratings are required to begin. Optimal stability is achieved with 200 ratings, though heavy users can contribute thousands of ratings for greater precision.

The image ranking process involves selecting preferences from random image pairs, allowing the AI to learn individual tastes and styles. Users can access the personalization tool at midjourney.com/personalize.

Organizational and Collaborative Improvements

The updates also enhance team collaboration. Founder David Holz emphasized the goal of making personalization more accessible and effective for a wider range of workflows, enabling users to balance creative control with team collaboration.

Earlier this week, MidJourney teased another experimental feature, Patchwork, a collaborative whiteboard for image creation. These recent updates signal MidJourney's ongoing commitment to refining its tools for both individual creators and collaborative teams.

What This Means

MidJourney’s latest features mark a significant step forward in empowering creators with more intuitive and customizable tools. Moodboards make it easier for users to visually articulate their creative visions, while personalization profiles ensure flexibility across multiple projects.

The reduced barrier to training custom models also opens the door for more users to leverage MidJourney's AI capabilities. These updates reinforce MidJourney’s position as a leader in AI-assisted art, offering powerful solutions for diverse creative needs.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing, image, and idea-generation support from ChatGPT, an AI assistant. However, the final perspective and editorial choices are solely Alicia Shapiro’s. Special thanks to ChatGPT for assistance with research and editorial support in crafting this article.