How to Make an Immunology Figure with AI: Step-by-Step Guide (2026)
A step-by-step how-to for making publication-ready immunology figures with AI — from your first prompt to export. Includes a four-step workflow, worked examples by field (vaccinology, oncology, autoimmunity), and tips for thesis and paper figures.
You need an immunology figure for a thesis chapter, a grant, or a manuscript — and you need it to look clean, labeled, and correct. This guide is the step-by-step how-to: not a list of prompts to copy, but the actual workflow you follow from a blank canvas to an export-ready file. You'll learn how to plan a figure, generate it, refine it conversationally, and ship it — with worked examples drawn from real subfields. If you just want prompts to copy, see our companion post, 25 immunology diagram prompts; this guide teaches you the method behind them.
By the end of this guide you'll be able to:
Make an immunology figure with AI in four repeatable steps — plan, generate, refine, export.
Adapt the workflow to your own field, whether that's vaccinology, tumor immunology, or autoimmunity.
Produce thesis- and paper-ready figures with correct labels, consistent arrows, and a clean layout.
Fix the issues that make AI immunology illustrations look wrong, before they reach a reviewer.
The whole workflow happens in the SciDraw AI editor, the AI-native workspace where your figure is the canvas and plain language is the tool.
Every good immunology figure — antibody, synapse, checkpoint, or cascade — comes out of the same four steps. Learn them once and you can draw anything in the field.
Plan the scene. Decide the subject, the cast (cells, receptors, molecules), and the relationships (who activates or inhibits whom). Write these down before you touch the editor.
Generate a first draft. Turn your plan into one descriptive sentence that names the subject, the interactions, the labels, and the style.
Refine conversationally. Don't restart — ask the AI to recolor, relabel, simplify, or rearrange the existing image.
Davie Chen is a researcher at the Faculty of Animation and Intermedia, University of Arts in Poznan, studying generative AI for scientific figure creation, patent illustration, and manuscript drafting. SciDraw AI is one of the research-to-product tools built from this work.
Thousands of researchers use SciDraw AI to make publication-ready figures for papers, grants, and journal submissions — in minutes, with no design skills.
Export and place. Download a high-resolution image or editable SVG/PPTX and drop it into your manuscript, slides, or poster.
Template for step 2: "Draw [subject] showing [interactions]. Label [cells/receptors/molecules]. Use a clean flat-vector style with activation arrows and inhibition bars."
The biggest difference between a figure that works and one that wastes credits is planning. Before prompting, answer three questions:
What is the single message? A figure should make one point. "PD-1 blockade restores T-cell killing" is a message; "everything about checkpoints" is not.
Who is in the scene? List the cells (dendritic cell, CD4+ T cell, tumor cell), the receptors (TCR, MHC II, PD-1, PD-L1), and the molecules (peptide, cytokines).
What are the relationships? Mark each as activation (arrow) or inhibition (flat-end bar), and note the direction of flow.
This three-part plan becomes the skeleton of your prompt. Naming each element is what tells the AI exactly what to label — the single most important habit for accurate immunology figures.
Turn the plan into one sentence. For an antigen presentation figure, the plan ("dendritic cell shows a peptide on MHC II to a CD4+ T cell; label APC, MHC II, peptide, TCR, CD4, CD28-B7") becomes a clean prompt and a clean result.
Tips that make the first draft land:
Name every element you want labeled. The AI labels what you name.
Specify direction. Use "activation arrow" and "inhibition bar" so immune signaling reads correctly.
Ask for a layout. "Two-cell synapse view" or "left-to-right flow" keeps dense scenes organized.
Keep in-figure text short. Put detail in the caption, not inside the diagram.
A first draft is rarely final, and that's fine. The AI-native advantage is that you edit by describing the change instead of redrawing. Common refinements:
"Recolor the CD4+ T cell blue and the tumor cell gray."
"The label 'PDL1' should read 'PD-L1'."
"Add an anti-PD-1 antibody blocking the PD-1/PD-L1 interaction."
"Simplify to the three main steps and remove the legend."
Each instruction edits the existing figure in place, so you never lose the parts you already like. For a deeper dive into label-only edits, see how to edit text and labels in an AI figure.
Message: how immunity forms after vaccination. Cast: antigen, APC, helper T cell, B cell, antibodies, memory cells. Flow: left to right. Prompt the antigen uptake, T- and B-cell activation, antibody production, and memory formation — then refine to highlight the memory step if that's your point. For mRNA vaccines, add "lipid nanoparticle delivery and spike-antigen translation" to the plan.
Message: how a therapy releases the immune brake. Cast: T cell, tumor cell, PD-1, PD-L1, anti-PD-1 antibody. Relationships: PD-1–PD-L1 inhibition (bar), blockade restoring killing (arrow). This is the checkpoint figure above; for CAR-T, replan the cast as an engineered T cell, a chimeric antigen receptor, and a tumor antigen in a left-to-right flow.
Message: how a cytokine cascade drives tissue inflammation. Cast: macrophage, TLRs, pathogen or self-antigen, TNF-α, IL-6, IL-1β, recruited neutrophils. Plan the recognition step, the cytokine release, and the recruitment, then refine colors to group pro-inflammatory mediators.
Thesis and journal figures have extra constraints. Build them in:
Match the journal's column width. Plan a single-column figure as a tall, simple layout; a double-column figure can be a multi-panel synapse-plus-cascade view. Generate at high resolution.
Use panel letters. Ask the AI to "add a panel letter 'A' top-left" so the figure slots into a composite.
Keep labels in the figure minimal. Reviewers want the detail in the legend. For help writing it, our companion prompts post lists the elements to name.
Stay consistent across chapters. Reuse the same color code and arrow style for every immunology figure in the document so the thesis reads as one voice.
When the figure is right, export it to editable SVG or PowerPoint (PPTX), or download a high-resolution image for your manuscript, slides, or poster. Need a colorblind-safe palette for the published version? See how to recolor a scientific diagram. Want a library of ready-made starting points instead of building from scratch? Browse the 25 immunology diagram prompts and drop any of them into the editor.
How do I make an immunology figure with AI?
Plan the scene (subject, cells, relationships), describe it in one sentence in the Immunology Diagram Generator, refine the result conversationally, then export. This four-step workflow produces publication-ready figures without design software.
What is the best way to draw an immunology diagram for a thesis?
Plan a single message per figure, generate at high resolution to match your journal's column width, add panel letters for composites, and reuse one consistent color and arrow style across every chapter so the thesis reads cohesively.
Can I create immunology figures for a paper for free?
Yes — you can start generating immunology figures for free in the SciDraw AI editor, then upgrade for more credits and editable SVG/PPTX export for your manuscript.
How do I fix a wrong label in an AI immunology figure?
Don't redraw — quote the current text and the correction, such as "change 'PDL1' to 'PD-L1'," and the AI edits only that label while keeping the layout, colors, and arrows intact.
Is an AI immunology figure generator a good BioRender alternative?
If you'd rather describe a figure than drag icons onto a canvas, an AI generator is a fast, low-cost alternative for antibody, antigen-presentation, T-cell, checkpoint, and vaccine diagrams, with editable export for collaborators.
Open the Immunology Diagram Generator, write your three-part plan, and turn it into a figure in the SciDraw AI editor. From antibody structures to checkpoint blockade, your next thesis or paper figure is one workflow away.
