Tech & AIBeginnerPreview
Midjourney for Art & Visual Storytelling
A hands-on course that teaches you to operate Midjourney as a reliable visual production tool, not a slot machine. You leave with a prompt formula, a parameter cheat sheet, reusable style references, and a workflow that keeps a whole campaign on-brand.
For marketers, founders, illustrators, and content creators who want to produce on-brand images with Midjourney without a design or coding background.
Course content
Workbook & downloads
Put the course into practice — a printable workbook plus editable templates you can fill in and reuse.
Preview the workbook
This workbook turns the course into a working studio system: a prompt library, a parameter cheat sheet, a locked style reference, a character sheet, and a moodboard-to-delivery pipeline you can hand to a client. Work through one section per module, running every prompt in Midjourney as you go and saving the results. By the end you will have a reusable set of presets and a finished, on-brand image set.
Getting Started with Midjourney
Set up your account, learn the action buttons, and run the generate, upscale, vary loop until it is second nature.
Exercise: Run the Core Loop
Generate one grid, then take a single image all the way to a refined result using only the action buttons. Do this three times with different subjects so the loop becomes automatic. Save the prompt and the best image from each run.
- Generate this prompt and read all four images before choosing: a cozy independent coffee shop interior at golden hour, warm light through large windows, wooden tables, plants, photographic, 35mm --ar 3:2
- Upscale your favorite of the four, then click Vary Subtle to produce close alternatives.
- Reroll the original prompt once and note how different the new grid is from the first.
Worksheet: Account and Setup Decisions
Fill this in once so you know your plan limits and where you will work. Revisit it if your monthly image volume changes, since unlimited Relax mode is the main reason to move up to Standard.
- Chosen plan (Basic, Standard, Pro, Mega) and monthly cost
- Approximate images you generate per month
- Do you need unlimited Relax mode? (yes or no)
- Do you need Stealth mode for private work? (yes or no)
- Primary interface you will use (Discord or web app)
- Default model version (7, 6.1, or Niji)
- Where you will save prompts and styles (folder or notes app)
Checklist: First-Run Setup Checklist
- Discord account created and linked to midjourney.com
- Active subscription confirmed under Manage Subscription
- Generated a few images so the web app unlocks fully
- Ran /settings and confirmed model version and Fast vs Relax mode
- Created a folder or notes file to start your prompt library
Prompt Craft and Parameters
Build prompts with the six-part formula and lock in the parameters that actually change your output.
Worksheet: Six-Part Prompt Builder
Fill all six parts for one real image you need, in order, then assemble them into a single prompt with the subject first. Reuse this sheet as a template for every new image so nothing important is left to chance.
- Subject and key traits
- Environment (where it is)
- Composition and shot (framing, angle, depth of field)
- Lighting (quality and direction)
- Medium and style (photo, painting, illustration, plus look)
- Mood and finishing detail
- Assembled prompt with parameters appended
Exercise: Parameter A/B Tests
Take one prompt and change a single parameter at a time so you can feel what each one does. Keep every other word identical. Note which setting you would use for brand work versus concept art.
- Run the same prompt three times at --stylize 50, --stylize 250, and --stylize 750 and compare how literal versus artistic each is.
- Run it again at --chaos 0 versus --chaos 50 and note how much the four images diverge.
- Add --weird 400 to one run and describe how the aesthetic shifts compared to the default.
Exercise: Negative and Multi-Prompt Practice
Practice steering the model away from unwanted elements and fusing two ideas on purpose. Save any combination that produces a clean, intentional result.
- Generate a scene, then add --no text, watermark, signature, extra fingers and compare the cleanliness of the result.
- Write a multi-prompt that biases one idea over another, for example: ancient library::2 greenhouse::1
- Combine high chaos and weird to explore, then re-run the winner at --chaos 0 to converge.
Checklist: Prompt Quality Check
- Subject is first and most specific in the prompt
- All six parts are present or deliberately omitted
- Filler words like very, beautiful, and amazing are removed
- Aspect ratio is set to match the final placement
- Stylize, chaos, and quality are set on purpose, not left to chance
Consistency and Style Control
Lock a single look across many images using style references, character references, and saved profiles.
Exercise: Lock a Style with --sref
Find a look you love, capture it as a style reference, then prove it holds by generating three completely different subjects that all share the look. This is the core skill for any branded set.
- Generate exploratory images until one nails your target look, then note its --sref code or image URL.
- Generate three different subjects (a person, a product, a landscape) each using that same --sref so they match.
- Adjust --sw to 50 and then 500 on one subject and note how much the reference dominates.
Exercise: Build a Consistent Character
Create one invented character and place them in three different scenes that still read as the same person. Generate a clean front-facing portrait first to use as your reference image.
- Generate a neutral, well-lit, front-facing character portrait to serve as your --cref source.
- Put that character in three new scenes using --cref [URL] --cw 0 to keep the face but change wardrobe and setting.
- Add your locked --sref so the character and the art style stay consistent across all three.
Worksheet: Style Library Entry
Document one signature look so anyone can reproduce it next quarter. Create one entry per recurring style or client. A look you cannot reproduce is a happy accident; this sheet makes it an asset.
- Style name (e.g. brand-hero-photoreal)
- Intended use
- --sref code or reference image URL
- --sw value
- Default aspect ratio
- Standard --no exclusions
- Two example image filenames
Checklist: Consistency QA
- Every image in the set uses the same --sref code or reference
- Style weight (--sw) is standardized across the set
- Recurring characters use --cref with a consistent --cw
- Aspect ratio and --no exclusions match across the set
- The recipe is documented so a teammate could reproduce it
Refining, Editing, and Shipping Brand Work
Take a promising image to a finished, rights-cleared deliverable with upscaling, inpainting, and the delivery pipeline.
Exercise: One Hero, Every Placement
Take one strong upscaled image and reframe it for three formats without regenerating from scratch. Use zoom and pan to create space rather than cropping away your composition.
- Use Custom Zoom to turn your hero image into a 16:9 banner so the model paints the sides.
- Pan up or Zoom Out to add headroom for a logo or headline.
- Export a 1:1, a 9:16, and a 16:9 version that all keep the same subject and style.
Exercise: Inpaint a Fix
Find an image that is almost perfect and repair the broken part with Vary Region instead of rerolling. Change one region at a time so each fix is controllable.
- Open an upscaled image in the Editor and mask a flaw (a hand, a stray object, or an off-brand color), masking slightly larger than the flaw.
- Edit the prompt to describe only what belongs in that region and regenerate just the mask.
- Leave a clean area for text and add the real headline afterward in Canva, Figma, or Photoshop.
Worksheet: Project Pipeline Tracker
Fill this for one real project to run the full moodboard-to-delivery pipeline. Following the same steps every time is what makes your output predictable and on-brand.
- Project and client
- Moodboard reference count and source
- Locked --sref code and --sw
- Aspect ratios needed for delivery
- List of assets to produce (subjects)
- Refinement notes (upscale, inpaint, reframe)
- Final export specs (dimensions and format)
Checklist: Pre-Delivery and Rights Check
- Keepers upscaled to the resolution the placement or printer requires
- Flaws fixed with inpainting; real text and logos added in a design tool
- Each placement size produced by zoom or pan, not destructive cropping
- Plan tier meets the commercial-use and revenue requirements in the current Terms
- No misleading deepfakes or living-artist name mimicry; disclosure added where expected
Your Action Plan
- Choose your plan, link Discord and the web app, and run the generate, upscale, vary loop three times
- Build the six-part prompt template and save five working prompts to your library
- Run the parameter A/B tests until you know what --stylize, --chaos, --weird, and --quality each do
- Explore until you find a signature look, then lock it as a saved --sref with a set --sw
- Create a reference character and place them in three consistent scenes using --cref and --sref
- Document each signature look in your style library so it is reproducible
- Produce a full asset set by swapping only the subject while keeping the locked preset
- Upscale keepers, inpaint flaws with Vary Region, and reframe each placement with zoom and pan
- Add real text and logos in Canva or Photoshop and export to the exact spec requested
- Confirm licensing and disclosure requirements before delivering any paid client work
Pairs well with
Courses members commonly take alongside this one.
Flagship CoursePreview
Freelance Business Foundations: Position, Price, Sell, and Deliver High-Value Services
Freelancing · Beginner · 16h
Self-pacedPreview
Client GrowthPreview
Freelance Client Acquisition: Outreach, Leads, Referrals, and Deal Flow
Freelancing · Beginner · 15h 30m
Self-pacedPreview
Sales SystemPreview
Freelance Sales & Proposals: Discovery Calls, Scoping, Objections, and Closing
Freelancing · Intermediate · 16h
Self-pacedPreview