Logo prompts. Wordmark, monogram, mark.
Pick a brand archetype, logo form, style era, color treatment, and palette mood; get a Midjourney, Nano Banana, or DALL-E ready prompt with vector-trace directive and explicit negative prompts. Browser-only; nothing is saved.
Brand, form, era, color.
Your logo prompt.
Click build to generate.
What good looks like.
Emani: clean-beauty wordmark in editorial serif, monochrome ink on cream
Form: wordmark only. Era: editorial serif (Tiempos Headline weight). Color: monochrome ink (#0f1713) on cream paper. The wordmark sits at 3:1 aspect, all-lowercase, generous tracking. Generated in Midjourney v6.1 at 4 variations, then traced as SVG via Adobe Illustrator's image-trace function. Pairs with cream-based packaging and editorial photography.
Linear: B2B SaaS monogram, geometric modern, duotone navy + electric green
Form: monogram (single L). Era: geometric modern (Futura proportions). Color: duotone (navy #0a1c3d, electric green #00ff88). The L is precision-engineered with right-angle terminals; reads the same at 16 px favicon and 1000 px header. Generated in Midjourney at 9 variations, top 2 refined in Figma.
Mother: agency abstract mark, hand-drawn ink, black-only on white
Form: abstract mark only (no letters). Era: hand-drawn organic (single ink-pen weight, intentional wobble). Color: black on white only - no accent. The mark feels made-by-hand which signals craft positioning to creative directors. Generated in Anthropic tooling or DALL-E 3 for hand-drawn fidelity, then redrawn in Illustrator as a clean vector.
A logo is a silhouette, not an illustration.
A logo ships at sizes from 16 px (browser favicon) to 1000 px (storefront signage). The 16 px size is what determines whether anyone recognizes your brand in a browser tab. Most AI-generated logo attempts fail because the prompt asked for an "illustration" (a tiny scene, a clever metaphor with three elements, a small story) rather than a silhouette. A good prompt names the form as one shape, fixes the era of typography, names the color, and adds the scalability directive: "must be readable from 16 to 1000 pixels." Without that line, the model packs detail it should have left out.
Form first, ornament never
The single biggest decision in a logo brief is the form: wordmark, monogram, abstract mark, pictorial mark, combination, or emblem. Each carries different production constraints. Wordmarks need typography you can defend at every size; monograms need a letterform that does not look like a placeholder; abstract marks need a shape distinctive enough to memorize; pictorial marks need a recognizable object that does not date. Most AI-generated logos fail because the prompt did not commit to a form — the model invented an over-decorated combination of all three. Pick one form per generation; iterate on the strongest before bringing in a second form.
Era of typography is the brand archetype
Editorial serif (Tiempos, Recoleta, GT Sectra) signals magazine-led DTC. Modern grotesque (Inter, Söhne, Neue Haas Grotesk) signals B2B SaaS engineering precision. Geometric modern (Futura, Avenir Next) signals Bauhaus-inflected design discipline. Hand-drawn signals craft and warmth. Vintage classic signals heritage. Brutalist signals art-school confidence. Art Deco signals luxury and 1920s premium. The era of the typography IS the brand archetype — pick the era that matches the audience the brand is selling to, not the era you personally find most beautiful.
Scalability is the contract
The single most important constraint on any logo is that it must remain identifiable at the smallest size the brand will use it at. For most brands that is the favicon (16 px to 32 px). Test by generating the logo at 1024 px and then scaling it down to 16 px in Figma. If the silhouette becomes mush, the form has too much detail. The fix is to drop ornament: thinner strokes, simpler curves, fewer elements. The scalability directive in the prompt — "must be readable from 16 to 1000 pixels" — forces the model to make those simplifications upfront.
Color is the second decision, not the first
The default AI-generated palette is too saturated and too rainbow. Real brand identity work starts in monochrome: prove the form works in pure black before adding color. Once the form is locked, the color treatment (monochrome, duotone, brand-led, muted, black-only) and the palette mood (warm earthy, cool corporate, saturated, luxury restrained, natural organic, mono-confident) become the second pass. Ask the AI for monochrome variations first; add color in a second generation only when the form is settled. This is the canonical approach taught by every reputable design school and reinforced in design-systems documentation.
Vector trace is mandatory
AI-generated logos are raster images. They blur at any non-native size and cannot be edited cleanly in Illustrator or Figma. After generation, run an SVG vector trace (Illustrator's Image Trace, Figma's Vectorize plugin, or the open-source Inkscape's Trace Bitmap). The trace converts the bitmap into clean vector paths; you can then refine the curves, adjust spacing, and export at any size without quality loss. A logo that ships only as a PNG is a logo that will look broken on a billboard.
Iterate at the cheapest fidelity
The biggest mistake is generating one logo at high fidelity and then trying to redesign it. Better: generate 9 to 16 variations at the start, eliminate the bottom 80% in 30 seconds (silhouette test, scale test, competitor-side-by-side), then take the top 2 to 3 for high-fidelity refinement in vector tools. The cheapest fidelity is the prompt itself — if you do not like the first batch, change the form or the era and regenerate; do not try to refine a bad form into a good one. This is the same iterate-at-low-fidelity approach taught in every reputable design program and reinforced by Stability AI's design guidance.
Related work: Brand identity for full brand systems beyond the logo. Logo design for the production-ready service. App icon prompt generator for the app-store counterpart. Favicon generator to ship the multi-size browser tab assets. Business name generator if the brand still needs naming.
Five answers.
Should I use Midjourney, Nano Banana, or DALL-E for logos?
Each handles a different form well. Midjourney v6.1 is strongest for wordmarks, geometric monograms, and modern grotesque typography because it preserves letterform precision. Nano Banana (Gemini's image model) is currently strongest for hand-drawn, pictorial, and editorial-serif logos because it preserves organic curves. DALL-E 3 is the most flexible for vintage and emblem styles. None of the three reliably produces production-ready vector output — always run an SVG vector trace on the winning raster before shipping.
Why does the prompt include a vector-trace directive?
All AI image generators output raster (pixel) images. A logo must ship as a vector (SVG) so it scales infinitely. The vector-trace directive in the prompt asks the model for clean curves and high contrast — both make the SVG conversion cleaner. After generation, use Illustrator's Image Trace, Figma's Vectorize plugin, or Inkscape's Trace Bitmap to convert the raster to SVG. Then refine the curves manually for production.
What's the difference between a wordmark, monogram, and mark?
A wordmark is the brand name set in chosen typography with no symbol — Google's logo, Coca-Cola's logo, FedEx's logo. A monogram is one or two letters from the brand name forming the entire mark — IBM's, HBO's, NASA's. An abstract mark is a distinctive shape with no letters — Nike's swoosh, Apple's apple, Twitter's bird. A combination mark uses both a graphic mark and a wordmark together — Adidas's three stripes plus the wordmark. Pick by recognizability needs: wordmarks teach the brand name, marks become memorable shapes once known.
Do I need a designer if the AI gives me good output?
For early validation, MVP launches, and pre-funding founder work, the AI output plus a vector trace is enough to ship. For production brand identity at scale, you need a designer to refine the curves, build the full system (color tokens, typography pairings, logo lockups, monogram + wordmark variations, social-media variants, monochrome + reverse + inline locks), and ensure the logo works across every touchpoint. The AI gives you the form; a designer ships the system. See our brand identity service for the full system spec.
Does this tool save my logo prompts?
No. Every value you enter and every prompt that gets built lives in your browser tab only. Nothing is transmitted to a server, stored in a database, or synced across devices. Close the tab and the data is gone. The Copy prompt button puts the assembled prompt on your clipboard; that is the only output path.
Logos are silhouettes.
Our brand-identity engagements take the AI-generated direction and build the full production system: vector master, color tokens, typography pairings, logo lockup variations, social variants, and brand book. Two-week turnaround on the logo system.
Published .