Midjourney Case Study: How an Indie Game Studio Created 200 Consistent Character Assets with Style References and Prompt Chaining

Background: The Consistency Challenge in Indie Game Art

When indie studio Ember Forge Games (a composite based on real workflows) began production on their narrative RPG featuring 40 unique characters — each requiring front, back, expression, and action poses — they faced a daunting pipeline: 200+ individual concept art assets that needed to look like they came from a single artist’s hand. With a two-person art team and a four-month deadline, they turned to Midjourney as a concept art accelerator. This case study breaks down the exact workflow, prompts, and techniques they used to achieve visual consistency at scale — a process any indie team can replicate.

Phase 1: Establishing the Visual Foundation

Step 1 — Create the Style Anchor Image

The team began by generating a single hero illustration that would define the entire project’s look. They iterated through approximately 30 generations before locking in their anchor. /imagine prompt: fantasy RPG character portrait, hand-painted illustration style, warm muted color palette, visible brushstrokes, soft ambient occlusion lighting, inspired by Eytan Zana and Sachin Teng, detailed armor with weathered leather textures —ar 2:3 —style raw —s 250 —v 6.1

Once they selected their anchor image, they upscaled it and saved the image URL for use as a style reference across every subsequent generation.

Step 2 — Lock the Style with —sref

Midjourney’s —sref (style reference) parameter became the backbone of the pipeline. By pointing every prompt to the anchor image, the team enforced a consistent color palette, rendering technique, and mood. /imagine prompt: female elven ranger, leather armor, short silver hair, confident stance, full body character concept art, neutral background —sref https://cdn.midjourney.com/YOUR_ANCHOR_IMAGE_ID/0_0.png —sw 100 —ar 2:3 —style raw —s 200 —v 6.1

Key parameters explained:

  • —sref [URL] — Points to the style anchor image- —sw 100 — Style weight at maximum to enforce strict adherence (range 0–1000, default 100)- —style raw — Reduces Midjourney’s default beautification for more controllable output- —s 200 — Moderate stylization for a balance of consistency and detail

Phase 2: Character Sheet Workflow

Step 3 — Generate Turnaround Sheets

For each character, the team generated a multi-view turnaround sheet to hand off to their 3D modeler. This required careful prompt engineering to get front, side, and back views in a single image. /imagine prompt: character design sheet, full body turnaround, front view side view back view, male dwarf blacksmith, leather apron, thick braided beard, heavy boots, fantasy RPG style, white background, clean reference sheet layout —sref https://cdn.midjourney.com/YOUR_ANCHOR_IMAGE_ID/0_0.png —sw 85 —ar 16:9 —style raw —v 6.1

Step 4 — Expression and Emotion Sheets

For key story characters, they generated facial expression sheets to guide the narrative team and 2D animators. /imagine prompt: facial expression sheet, 6 emotions, angry happy sad surprised determined fearful, female elven ranger, silver hair, close-up portrait grid, fantasy RPG character, white background --sref https://cdn.midjourney.com/YOUR_ANCHOR_IMAGE_ID/0_0.png --sw 90 --cref https://cdn.midjourney.com/YOUR_RANGER_IMAGE_ID/0_0.png --cw 100 --ar 16:9 --v 6.1

The --cref (character reference) parameter was critical here — it told Midjourney to maintain the same character's facial features and identity, while --sref maintained the art style.

Phase 3: Prompt Chaining for Scaled Production

Step 5 — Build a Prompt Template System

The team created a structured prompt template and stored it in a shared spreadsheet to ensure every artist and every batch followed the same formula. TEMPLATE: /imagine prompt: [VIEW_TYPE], [CHARACTER_DESC], [OUTFIT_DETAILS], [POSE_ACTION], fantasy RPG character concept art, [BACKGROUND] —sref [ANCHOR_URL] —sw [STYLE_WEIGHT] —cref [CHARACTER_URL] —cw [CHAR_WEIGHT] —ar [RATIO] —style raw —v 6.1

EXAMPLE FILL: /imagine prompt: full body action pose, male dwarf blacksmith, leather apron and heavy gauntlets, swinging a warhammer mid-strike, fantasy RPG character concept art, simple gradient background —sref https://cdn.midjourney.com/ANCHOR_ID/0_0.png —sw 90 —cref https://cdn.midjourney.com/DWARF_ID/0_0.png —cw 80 —ar 2:3 —style raw —v 6.1

Step 6 — Batch Processing with Repeat

For generating variations efficiently, they leveraged Midjourney's --repeat parameter (available on Pro plans) to queue multiple generations from a single command. /imagine prompt: full body character concept art, female human merchant, layered traveling robes, carrying a ledger, standing pose, neutral background --sref https://cdn.midjourney.com/ANCHOR_ID/0_0.png --sw 90 --style raw --v 6.1 --repeat 4

This generated four independent batches (16 images total), dramatically accelerating the selection process.

Results: By the Numbers

MetricTraditional PipelineMidjourney-Assisted Pipeline
Total assets produced~50 in 4 months200+ in 4 months
Time per character sheet8–12 hours1.5–3 hours (generation + curation + touchup)
Style consistency rating (internal review)85%93%
Cost (art labor)$24,000$6,200 + $120 Midjourney Pro annual
Revision rounds per character3–51–2 (select from variations)
## Pro Tips for Power Users - **Combine --sref and --cref together** — Use style reference for art direction and character reference for identity. Adjust --sw and --cw independently to fine-tune the balance.- **Use multiple style references** — You can pass up to three URLs to --sref separated by spaces to blend influences: --sref URL1 URL2 --sw 60- **Save --sref codes with /shorten** — Use /shorten to analyze your prompts and remove unnecessary tokens that introduce style drift.- **Pin your seed for iterative refinement** — When adjusting a single character, add --seed 12345 to reduce variation between runs while you tweak prompt language.- **Export a master parameter block** — Store your --sref, --sw, --style, --s, and --v values in a shared document so every team member uses identical settings.- **Use --no to suppress unwanted elements** — If backgrounds keep appearing too detailed, add --no detailed background, gradient, scenery to keep the focus on the character. ## Troubleshooting Common Issues
ProblemCauseSolution
Characters look different despite --srefStyle weight too low or prompt language overriding styleIncrease --sw to 100–200. Remove conflicting style descriptors from prompt text (e.g., "watercolor", "cel-shaded")
Turnaround sheets show inconsistent posesMidjourney interprets multi-view looselyAdd explicit spatial cues: "front view on left, side view center, back view on right". Use --ar 16:9 or wider
--cref produces a different faceCharacter weight too low or source image too complexIncrease --cw to 100. Use a clean, well-lit headshot as the --cref source rather than a full scene
Colors drift across batchesDifferent background prompts influence color gradingStandardize background language ("flat white background" or "neutral gray background") and keep --style raw active
"Job queued" errors on --repeatFree or Basic plans don't support --repeatUpgrade to Pro or Mega plan. Alternatively, submit prompts manually in smaller batches
## Key Takeaways - **Anchor early** — Invest time in generating one perfect style anchor image before any production work begins.- **Systematize prompts** — Template-driven prompts eliminate drift far better than ad-hoc writing.- **Layer references** — --sref for style, --cref for character identity, --seed for iteration stability.- **Midjourney is a concept accelerator, not a replacement** — The studio still spent 30% of art time on manual touchups, color correction, and detail painting in Photoshop. ## Frequently Asked Questions

Can Midjourney fully replace a character artist for indie game production?

No. Midjourney excels at rapid concept exploration and establishing consistent visual direction, but production-ready game assets still require manual refinement. In this case study, the studio used Midjourney output as a 70% starting point, then completed detail painting, color correction, and asset formatting in Photoshop and Clip Studio Paint. The value lies in compressing the ideation and concepting phase from days to hours.

How many style reference images should I use for maximum consistency?

Start with one strong anchor image and use it as your sole —sref for the first 20–30 assets. Once you’ve validated that the style holds, you can experiment with adding a second reference to introduce controlled variation (e.g., a reference for lighting mood). Using more than two references tends to dilute consistency. Keep —sw between 85–100 for strict adherence during production runs.

What Midjourney plan do I need for a production workflow like this?

The Pro plan ($96/month or $576/year) is the minimum recommended tier. You need access to —repeat for batch processing, Stealth Mode to keep proprietary character designs private, and sufficient Fast GPU hours for iterative refinement. For studios generating 200+ assets, the Mega plan provides more Fast hours and avoids reliance on Relax mode, which can slow production timelines significantly.

Explore More Tools

Grok Best Practices for Academic Research and Literature Discovery: Leveraging X/Twitter for Scholarly Intelligence Best Practices Grok Best Practices for Content Strategy: Identify Trending Topics Before They Peak and Create Content That Captures Demand Best Practices Grok Case Study: How a DTC Beauty Brand Used Real-Time Social Listening to Save Their Product Launch Case Study Grok Case Study: How a Pharma Company Tracked Patient Sentiment During a Drug Launch and Caught a Safety Signal 48 Hours Before the FDA Case Study Grok Case Study: How a Disaster Relief Nonprofit Used Real-Time X/Twitter Monitoring to Coordinate Emergency Response 3x Faster Case Study Grok Case Study: How a Political Campaign Used X/Twitter Sentiment Analysis to Reshape Messaging and Win a Swing District Case Study How to Use Grok for Competitive Intelligence: Track Product Launches, Pricing Changes, and Market Positioning in Real Time How-To Grok vs Perplexity vs ChatGPT Search for Real-Time Information: Which AI Search Tool Is Most Accurate in 2026? Comparison How to Use Grok for Crisis Communication Monitoring: Detect, Assess, and Respond to PR Emergencies in Real Time How-To How to Use Grok for Product Improvement: Extract Customer Feedback Signals from X/Twitter That Your Support Team Misses How-To How to Use Grok for Conference Live Monitoring: Extract Event Insights and Identify Networking Opportunities in Real Time How-To How to Use Grok for Influencer Marketing: Discover, Vet, and Track Influencer Partnerships Using Real X/Twitter Data How-To How to Use Grok for Job Market Analysis: Track Industry Hiring Trends, Layoff Signals, and Salary Discussions on X/Twitter How-To How to Use Grok for Investor Relations: Track Earnings Sentiment, Analyst Reactions, and Shareholder Concerns in Real Time How-To How to Use Grok for Recruitment and Talent Intelligence: Identifying Hiring Signals from X/Twitter Data How-To How to Use Grok for Startup Fundraising Intelligence: Track Investor Sentiment, VC Activity, and Funding Trends on X/Twitter How-To How to Use Grok for Regulatory Compliance Monitoring: Real-Time Policy Tracking Across Industries How-To NotebookLM Best Practices for Financial Analysts: Due Diligence, Investment Research & Risk Factor Analysis Across SEC Filings Best Practices NotebookLM Best Practices for Teachers: Build Curriculum-Aligned Lesson Plans, Study Guides, and Assessment Materials from Your Own Resources Best Practices NotebookLM Case Study: How an Insurance Company Built a Claims Processing Training System That Cut Errors by 35% Case Study