Pre-Viz in a Flash: A Filmmaker’s Guide to AI Storyboarding with Midjourney
Is AI going to replace the director? The cinematographer? The storyboard artist? The short answer is no. But as of July 9, 2025, a filmmaker who knows how to collaborate with AI will conceptualize, pitch, and execute their vision faster than ever before. Forget the existential dread. It’s time to meet your new pre-visualization supervisor—one with an encyclopedic knowledge of film history and an inexhaustible work ethic. Today, we’re not just making pretty pictures; we are building the visual blueprint for your next film.
The Director’s Dilemma: The Time and Cost of Vision
Every filmmaker knows the friction point between the vibrant movie playing in your head and the stark reality of communicating it. Storyboarding is essential, but it’s a bottleneck. It can be slow, expensive, and a frustrating game of telephone between you and the artist. What if you could generate an entire scene’s worth of keyframes, in a consistent style, in the time it takes to brew a coffee? This is not a fantasy. This is a new workflow.
We will use Midjourney, the premier AI image generator, not as a replacement for a skilled storyboard artist, but as a revolutionary tool for rapid ideation and visual discovery. You’ll go from a vague idea—’a detective finds a clue in a rainy sci-fi city’—to a tangible, editable, and shareable sequence of shots in minutes.
The Prompting Studio: The Establishing Shot
First, let’s set the scene. Every film needs an establishing shot that defines the world. We need to be specific with our language to tell the AI not just what to create, but how it should feel. Open Midjourney via Discord.
Copy and paste this prompt:
/imagine prompt: cinematic film still, a vast cyberpunk cityscape at night, inspired by Kowloon Walled City, dense with glowing holographic advertisements and exposed wiring, pouring acid rain, neo-noir aesthetic, shot on Arri Alexa with anamorphic lens, deep shadows and pockets of neon blue and magenta light –ar 2.39:1 –style raw –s 250
Press Enter. Midjourney will now generate four distinct options for your opening shot, each a potential universe to explore.
Strategist’s Log (Deconstructing the Prompt): We didn’t just say ‘sci-fi city’. We used ‘Kowloon Walled City’ as a historical and architectural anchor for density and chaos. ‘Neo-noir’ cues the mood and lighting. Mentioning ‘Arri Alexa’ and ‘anamorphic lens’ tells the AI we’re thinking like a cinematographer, pushing it towards a more professional, filmic look. Critically, –ar 2.39:1 forces the widescreen Cinemascope aspect ratio, instantly making it feel like a movie, not a social media post. –style raw gives us a more photographic and less ‘opinionated’ starting point, while –s 250 (stylize) increases the artistic interpretation.
The Consistency Challenge: Building a Cohesive Scene
You have your beautiful establishing shot. Great. But now you need a medium shot of your protagonist within that same world. This is where most creators get frustrated. They generate a new image, and the character looks different, the lighting is wrong, the style is off. The secret to cohesive storyboarding lies in creating a ‘Visual Prompt Locket’—a core set of descriptive keywords you’ll carry from shot to shot, and in using Midjourney’s more advanced features.
First, pick the establishing shot you like best from the initial grid and upscale it. Once you have the upscaled image, click ‘Vary (Subtle)’. This will give you the ‘seed’ number for that image, a key to its unique look. For advanced consistency, we’ll use Midjourney’s new Character Reference (`–cref`) feature. First, you need an image of your character. Let’s imagine you’ve already generated a character concept you like and have its URL.
The Prompting Studio: The Medium Shot
Now we push in on our character. We’ll reference our character’s image URL and reuse our ‘Visual Locket’ of keywords.
Copy and paste this prompt:
/imagine prompt: medium shot, a cynical private detective with a cybernetic eye, standing on a crowded market street, looking off-camera, drenched in pouring acid rain, neo-noir aesthetic, background is a dense cyberpunk city with neon signs –cref [URL_of_your_character_image.jpg] –cw 100 –ar 2.39:1 –style raw
This command tells Midjourney to create a new shot but to strongly base the character’s appearance on the image you provided.
Strategist’s Log (Maintaining Consistency): The –cref parameter is a game-changer. It takes the URL of an existing image and attempts to transfer the face, hair, and clothing of the character to the new generation. –cw 100 (Character Weight) sets the focus to maximum on the character’s features. We kept key phrases like ‘acid rain’, ‘neo-noir’, and ‘cyberpunk city’ to ensure the world remains consistent, even as we change the shot type from ‘wide’ to ‘medium’. This is how you build a believable sequence.
From Keyframes to Animatic: The Final Step
You can repeat this process for every shot in your scene: close-ups, inserts, reverse shots. For a close-up on an object, you’d prompt: “extreme close-up on a futuristic data-slate, glitching screen reflecting neon signs, held in a gloved hand, cinematic, neo-noir aesthetic…”. In under an hour, you can generate 20-30 high-quality, stylistically consistent storyboard panels.
The final step in this pre-visualization workflow is to create an animatic. This is a rough-cut of your film using the static storyboard images. Simply import your generated shots into any video editing software (Adobe Premiere, DaVinci Resolve, even iMovie). Sequence them in the timeline, adjust the timing for each shot to match your intended pacing, and add a temporary music track or sound effects. Suddenly, your vision is no longer just in your head. It’s a watchable, shareable asset you can use to pitch investors, align your crew, or simply refine your own directorial instincts.
The Big Questions: Your AI Debrief
“Is this going to put storyboard artists out of a job?”
No, it’s changing the job description. Think of this process as creating the ‘director’s pass’. You use AI to rapidly explore a hundred ideas on your own time. This saves immense amounts of time and money in the early stages. When you’re ready to commit, you bring these high-quality AI boards to a human storyboard artist not for creation, but for refinement. The artist can then add dynamic posing, fix inconsistencies, and infuse true human emotion into the keyframes. It elevates their role from ‘draftsman’ to ‘visual collaborator’.
“What about copyright? Can I use these images in my final film?”
This is the most critical question. According to Midjourney’s terms of service (as of this writing), paid subscribers own the assets they create. However, the legal landscape surrounding AI-generated art is still evolving, especially regarding copyright registration of purely AI output. For pre-visualization, mood boards, and pitching, you are generally in a very safe zone. If you plan to use an unedited AI image as a matte painting or a direct shot in a major commercial release, consulting with an entertainment lawyer is advisable. The ethical workflow is to use AI for ideation and then have human artists create the final, fully-owned assets based on the AI concepts.
“How do I develop a unique ‘AI Style’ that doesn’t look generic?”
Your unique style comes from the sum of your inputs. It’s the result of your specific taste in film, photography, and art. Instead of using generic terms like ‘sci-fi’, reference specific directors, cinematographers, or even painters. Try ‘in the style of Andrei Tarkovsky’, ‘cinematography by Roger Deakins’, or ‘with the color palette of a Zdzisław Beksiński painting’. Blend these influences. The AI is a mirror of the references you feed it. Your unique vision is expressed through the curation and combination of those references. The more specific and eclectic your prompts, the less ‘AI-ish’ your output will be.
Your Creative Sandbox Assignment
Your mission is to storyboard a simple, three-shot narrative sequence. The story: An archaeologist discovers a glowing artifact in a lost jungle temple.
- Shot 1 (Wide): Generate the establishing shot of the temple entrance, overgrown with ancient vines, with a single beam of light piercing the canopy. Focus on the atmosphere. Think ‘Indiana Jones meets Annihilation’.
- Shot 2 (Close-up): Generate a shot of the archaeologist’s face, illuminated by an otherworldly glow from below. Use the character consistency techniques we discussed. Capture a look of awe and fear.
- Shot 3 (Insert): Generate an extreme close-up of the artifact itself—a crystalline, intricate object humming with light. Make this prompt hyper-detailed.
Arrange these three images in sequence. You’ve just directed your first scene with an AI collaborator.
Your AI Integration Plan This Week
- Monday: Take the script for a project you’re working on. Choose one scene. Spend 20 minutes writing five different prompts for just the establishing shot, each with a different stylistic influence (e.g., ‘Kurosawa style’, ‘Wes Anderson symmetry’, ‘David Fincher’s color grading’).
- Wednesday: Select your favorite establishing shot. Now, generate a character concept sheet. Prompt for a ‘full body character design turnaround’ for your protagonist, making sure to describe their personality and clothing. Find the URL of the best one.
- Friday: Using your chosen establishing shot’s mood and the character’s `–cref` URL, generate the rest of the keyframes for the scene. Aim for at least five distinct shots (medium, close-up, over-the-shoulder, etc.).
- Sunday: Assemble your shots into a 30-second animatic. Add a temp track from a royalty-free music library. Watch it. You have now gone from page to screen, bridging the gap between imagination and execution.
Post Comment
You must be logged in to post a comment.