Image
AI Character Design on Martini
Indie game studio prototyping a 2D platformer needs the protagonist designed end-to-end — silhouette, color palette, expression sheet, outfit variations — before the rig artist starts. Drop the character description and style anchor, fan across Midjourney, Flux, Nano Banana 2, Flux Kontext, and GPT Image 2 to land the master portrait, then chain it into turnaround sheet, expression sheet, and outfit variations. The output is a character bible — concept stage for the artist team to clean.
What this feature solves
Character design is a multi-deliverable phase, not a one-image generation. An indie game studio building a 2D platformer protagonist needs a master portrait, a turnaround sheet (front, back, three-quarter), an expression sheet (happy, angry, surprised, sad, sleeping, determined), and outfit variations (default, alternate costume, holiday skin) — all anchored to the same character. Tab-based AI image tools generate one portrait per session and have no way to lock the design across the dozen-plus deliverables that constitute a real character bible.
The deeper problem is consistency at the design stage itself. A novelist writing a Substack series with recurring protagonists, an animator building a 2D-rigged short with one returning character, a tabletop publisher producing a campaign book with a fixed cast — every long-form creative project depends on the character holding identity across hundreds of generations. Without an upstream anchor that locks silhouette, color palette, and outfit signatures, the character drifts within the same project and the world breaks immersion.
And there is the legal honesty layer. Character design in the style of Pixar, Studio Ghibli, Marvel, Disney, or a living artist (Loish, Karla Ortiz, Greg Rutkowski) is exactly where AI image work runs into trademark and right-of-publicity risk. Style mimicry of a registered IP or a recently active artist is litigation-prone for commercial release. A workflow that frames AI character design as concept exploration — build an original style, anchor it, and frame the output as reference for the real character artist to clean — is honest about where AI fits in the production pipeline.
Why Martini is different
Martini treats character design as a multi-anchor canvas problem. The protagonist description lives as a text node — 'wiry teen runner, deep navy hood, copper-rimmed goggles, signature scar on the left cheek.' The style reference (line weight, color script, world tone) lives as another labeled image node. Wire both into Midjourney for the master portrait. Once the master lands, every downstream deliverable — turnaround, expression sheet, outfit variations — wires into the master portrait as the new upstream anchor. The character holds identity because the chain pulls from one locked source, not from a re-typed prompt.
Multi-model fanout for the master portrait, single-model for the deliverables. For the master, fan across Midjourney (editorial composition), Flux (high-fidelity rendering), Nano Banana 2 (reference fidelity for sequel work), Flux Kontext (edit-aware refinement), and GPT Image 2 (cleanup pass). Pick the strongest direction. For the turnaround, expression sheet, and outfit variations — the catalog of character views that make up the bible — lock the winning model and run them all from the same master portrait anchor. The bible scales on the chain rather than on per-view re-prompting.
Downstream chaining hands off to the production pipeline. Once the bible lands, chain into the image-upscale tool for high-resolution masters that the rig artist can clean and rig in Spine, Live2D, or DragonBones. Save the canvas as a character template; future sequels, season-two storylines, or alternate-universe variations inherit the locked anchors and only swap the per-view prompts. Frame the deliverable honestly — concept-stage character design for an artist team to clean and finalize, not engine-ready production sprites.
Common use cases
2D platformer protagonist for an indie game studio
Game studio designs the protagonist end-to-end — master portrait, turnaround, expression sheet, outfit variations — before the rig artist and animator start.
Character bible for a Substack or KDP serial novelist
Novelist produces character portraits for protagonists across a serial — same character holds across episode covers, marketing posts, and chapter-art reveals.
Expression sheet for a 2D-rigged animated short
Animator builds an expression sheet — happy, angry, surprised, sad, sleeping, determined — from the same character anchor for a short film rig.
Tabletop campaign cast for a Kickstarter book
Tabletop publisher designs the campaign cast — six characters at consistent style — for a Kickstarter campaign book and reference deck.
Brand mascot character bible for a recurring marketing role
Brand commissions a recurring mascot character with a multi-pose, multi-outfit bible for use across campaigns over the year.
Webtoon or manga protagonist with cast support
Indie webtoon creator designs the protagonist plus deuteragonist plus antagonist with consistent style for a long-form serial.
Recommended model stack
midjourney
imageEditorial composition and dramatic character poses for the master portrait that defines the bible.
flux
imageHigh-fidelity rendering for the polished master portrait and signature outfit variations.
nano-banana-2
imageReference-faithful re-generation that holds the character across the expression sheet and turnaround views.
flux-kontext
imageEdit-aware refinement for outfit swaps and pose variations from the master portrait anchor.
gpt-image-2
imageEdit-aware cleanup for the strongest character views before bible export.
How the workflow works in Martini
- 1
1. Write the character description as a text node
Capture silhouette, color palette, outfit signatures, age, mood, world context. The description lives as the upstream text anchor that every downstream view pulls from.
- 2
2. Anchor the style reference
Drop one style reference image as a labeled image node — line weight, color script, world tone. Build an original style; do not name a known IP or living artist.
- 3
3. Fan out the master portrait across multiple models
Run Midjourney, Flux, Nano Banana 2, Flux Kontext, and GPT Image 2 in parallel from the description and style anchors. Compare takes; pick the strongest master portrait.
- 4
4. Lock the master portrait as the new upstream anchor
The master portrait becomes the reference for every downstream view. Drop it as a labeled image node and replace the original description-only anchor with this richer visual source.
- 5
5. Generate the turnaround sheet
Wire the master portrait into Nano Banana 2 nodes for front, back, and three-quarter views. The turnaround locks the silhouette across angles.
- 6
6. Generate the expression and outfit sheets
Run the expression sheet (happy, angry, surprised, sad, sleeping, determined) and outfit variations (default, alternate, holiday) from the same master anchor through the locked model.
- 7
7. Save the canvas as a character template and hand off to the artist
Save the bible canvas. Future sequels or new storylines inherit the master anchor automatically. Hand the bible to the character artist for cleanup and rigging.
Example workflow
Lin is the art director at a three-person indie game studio building a 2D platformer about a teenage runner navigating a post-flood city. The protagonist is Mira — wiry, determined, deep navy hood, copper-rimmed goggles, signature scar on the left cheek. Lin opens a workspace canvas and writes the description as a text node, drops a style reference (clean line work, painterly background, muted urban palette), and fans the master portrait across Midjourney, Flux, Nano Banana 2, Flux Kontext, and GPT Image 2. The Midjourney take lands the strongest editorial pose. Lin locks Midjourney's output as the new master anchor and runs the turnaround through Nano Banana 2 — front, back, three-quarter. Then the expression sheet — Mira determined, exhausted, surprised, victorious, scared, asleep. Then outfit variations — default hood, alternate jumpsuit, holiday skin. The bible lands as a 12-view character canvas. Lin chains each view through GPT Image 2 for cleanup and image-upscale for high-res masters. The character artist on the team takes the bible into Procreate and Photoshop for final line cleanup and color polish before the rig artist takes over in Spine. The studio frames the deliverable honestly in the team handoff doc — concept-stage bible, not engine-ready sprite. Production cleanup happens after.
Tips and common mistakes
Tips
- Build an original style. Do not name Pixar, Studio Ghibli, Marvel, Disney, or a living artist in the prompts — style mimicry is legal-risk territory for commercial release.
- Anchor the master portrait once after fan-out. Every downstream view (turnaround, expression sheet, outfit variations) pulls from the master, not from a re-typed description.
- Use Nano Banana 2 for the catalog of character views. Reference-faithful generation holds the character across angles and emotions better than peers.
- Save the canvas as a character template. Sequel or season-two storylines inherit the master anchor automatically; you only swap the new view prompts.
- Frame the AI output honestly — concept-stage bible for the artist team to clean and finalize, not engine-ready sprite output.
Common mistakes
- Generating characters in the literal style of a known IP — Pixar, Studio Ghibli, Marvel, Disney — for commercial release. Style mimicry is legal-risk territory and exposes the team to litigation.
- Style mimicry of a living artist (Loish, Greg Rutkowski, Karla Ortiz) post-2023. Recent litigation has narrowed the safe zone considerably; build an original style or use generic visual cues instead.
- Skipping the master-portrait anchor. Without it, the turnaround drifts on silhouette, the expression sheet drifts on identity, and the bible falls apart visibly.
- Treating AI output as engine-ready. For commercial games, expect a character artist to clean and finalize the AI concept before rigging in Spine, Live2D, or 3D pipeline.
- Letting the supporting cast share an anchor with the protagonist. Each named character gets its own anchor; the world grows by adding anchors, not by re-prompting.
Related how-to guides
Related models and tools
Tool
AI Image Upscaling
Upscale images and keyframes before final video generation on Martini.
Tool
AI Background Removal
Remove backgrounds from images for assets and compositing on Martini.
Provider
Google's Veo video, Imagen image, and Nano Banana model workflows on Martini.
Provider
OpenAI
OpenAI's GPT Image and Sora video model workflows available on Martini.
Provider
ByteDance
ByteDance's Seedance video and Seedream image model families on Martini.
Related features
AI Character Reference — Reference-Image Workflows on Martini
Use reference images to guide AI model outputs on Martini's canvas.
AI Character Consistency Across Images and Video
Keep a subject consistent across image and video generations on Martini using reference workflows.
AI Game Asset Generator — Sprites, Concept Art, Backgrounds
Generate game-ready assets, sprites, concept art, and backgrounds on Martini.
AI Comic Strip Generator — Multi-Panel Comics on Martini
Generate multi-panel comic strips with consistent characters on Martini's canvas.
AI Photo Restoration — Restore Old Photos on Martini
Restore old, damaged, or low-quality photos with AI on Martini's canvas.
AI Style Transfer — Apply Artistic Styles to Images on Martini
Transfer artistic styles between images using AI on Martini.
AI Product Photography — Studio-Quality Product Images on Martini
Generate studio-quality product photos for e-commerce on Martini's canvas.
AI Headshot Generator — Professional Headshots in Minutes
Generate professional headshots for LinkedIn, resumes, and team pages on Martini's canvas.
AI Mockup Generator — Product, Device, and Brand Mockups
Generate product, device, and brand mockups for marketing on Martini's canvas.
AI Thumbnail Generator — YouTube and Social Thumbnails
Generate scroll-stopping thumbnails for YouTube, podcasts, and social on Martini.
AI Logo Generator — Brand Marks and Wordmarks on Martini
Generate logo concepts, brand marks, and wordmarks on Martini's canvas.
AI Emoji Generator — Custom Emoji on Martini
Generate custom emoji and stickers for Slack, Discord, and brand on Martini.
AI Sticker Generator — Telegram, WhatsApp, Discord Packs
Generate sticker packs for Telegram, WhatsApp, Discord, and iMessage on Martini.
AI Presentation Slides — Pitch Decks and Slide Visuals
Generate slide visuals, pitch deck imagery, and presentation graphics on Martini.
AI Icon Generator — App and UI Icons on Martini
Generate app icons, UI icons, and brand icon sets on Martini's canvas.
AI Architecture Rendering — Building and Space Visualization
Generate architectural renderings, exterior visualizations, and concept art on Martini.
AI Interior Design — Room and Space Visualization on Martini
Visualize interior designs, room concepts, and decor schemes on Martini's canvas.
Related docs
Related reading
Comparisons
Frequently asked questions
How is this different from ai-character-reference?
ai-character-reference is the primitive — how the reference node itself works on the canvas, the multi-anchor handling, the role-clear identity input. ai-character-design is the concept-design phase — turning a written description into a fully realized original character with master portrait, turnaround, expression sheet, and outfit variations. The reference page is the workflow tool; the design page is the deliverable phase that uses it.
How is this different from ai-character-consistency?
ai-character-consistency is the cross-modal outcome — the same character holding identity across images and videos in production work. ai-character-design is upstream of that — the original concept and bible build, before any production shots happen. Design produces the bible; consistency holds the bible across shots, scenes, and modalities.
Can I generate characters in the style of Pixar or Studio Ghibli?
For personal study and pure reference, image models will produce in those styles, but for commercial release the answer is no. Generating characters in the style of a known IP — Pixar, Studio Ghibli, Marvel, Disney — for products you ship is trademark and style-mimicry territory and creates real legal risk. Build an original style with mood-board sources rather than naming a registered brand or living artist.
Which model is best for character design?
For the master portrait, Midjourney leads on editorial composition and dramatic poses. Flux delivers high-fidelity rendering for polished portraits. For the catalog of character views (turnaround, expression sheet, outfit variations) Nano Banana 2 holds character identity across re-generations the strongest. Flux Kontext and GPT Image 2 refine the bible before export.
Will my character hold identity across all 12 bible views?
With a locked master-portrait anchor, yes — far better than re-typing the description into 12 sessions. Even so, expect some drift on small details (a button on the jumpsuit, the goggle strap) and review every view before declaring the bible complete. For commercial games and serial novels, the character artist usually does a final cleanup pass on the AI output before the bible ships to the rig artist.
Is the AI output engine-ready for a 2D rig?
No. AI character output is concept stage — it accelerates the bible build but does not produce engine-ready sprites, normal maps, UV unwraps, or rigged skeletons. For shipping titles, expect a character artist to clean and finalize the AI concept in Procreate or Photoshop, then hand to the rig artist for Spine, Live2D, or DragonBones rigging. Frame Martini as concept acceleration, not artist replacement.
Build it on the canvas
Open Martini and wire this workflow up in minutes. Free to start — no card required.