What this category actually does
You type a description, and it produces an original image — a photo-style shot, an illustration, a logo concept, product art. Nothing is copied from an existing photo; it’s generated fresh from the description, though the training data behind these tools is a real and unsettled conversation worth being aware of.
How it’s different from the categories next to it
This produces still images only. Video generation (a separate page in this directory) creates moving footage, and while some tools do both, they’re built and priced differently depending on which one you need. Avatars are a specific, narrower case — a digital person speaking — rather than open-ended image creation.
Unless noted otherwise, the tools below are covered based on research, testing, and independent comparisons — not day-to-day personal use. This directory combines firsthand experience with current research; where a tool has been used directly, that’s stated explicitly.
A note on training data and copyright: how these tools were trained, and what rights that gives you over the output, is still an active legal conversation across the industry, not a fully settled question. That’s part of why the commercial-license section below matters more in this category than in most others.
The real tools inside this category
Midjourney is worth choosing if raw visual quality and a distinctive artistic look matter most to you — it’s frequently cited in independent testing for strong aesthetic results, particularly for artistic or stylized work, and it’s a common choice for artists and designers for that reason. “Best-looking” is genuinely a matter of taste, so browsing a gallery of real output before subscribing is worth doing.
DALL-E (built into ChatGPT) is a reasonable starting point if you’re already using ChatGPT and want image generation without adding another subscription or learning another interface — it’s also generally considered easier to iterate with conversationally than the others here.
Adobe Firefly is worth a serious look specifically for commercial work — it was trained only on licensed and public-domain images, and Adobe offers commercial legal indemnification on Firefly output, a protection we haven’t seen matched elsewhere in this category. If you’re generating images for paid client work or ads, that protection is worth weighing against pure output quality.
Ideogram is the pick if you specifically need readable text inside the image — a poster, a sign, a label — where most other tools in this category still struggle to render legible text reliably.
What it typically costs
- Free plan: several tools here (Ideogram, and DALL-E via ChatGPT’s free tier or Bing’s free image creator) offer a genuine no-cost way to generate a limited number of images daily.
- Entry paid tier: commonly in the $10–15/month range for a starter credit allotment.
- Higher tiers: scale up from there, often $30+/month for heavier commercial use.
- Pricing model: typically credit-based — you get a monthly allotment of generations rather than unlimited use, even on paid plans.
Prices and credit amounts shift often — check current details before subscribing.
Good to know before you subscribe
Availability varies more here than in other categories: some tools are browser/web-only, others (like Midjourney) have historically required a specific interface (originally Discord-based, though this has evolved) — check current access details before assuming a standard web app. Commercial usage rights differ meaningfully by tool and plan; this is the category where it’s most worth reading the license terms before using an image in paid client or commercial work, not just assuming it’s covered. Export is typically standard image formats (PNG/JPG).
When you’d actually reach for this
Original artwork, product mockups, marketing visuals, blog or social graphics, or concept art — anywhere you need an image that doesn’t exist yet and don’t have (or want to hire) a photographer or illustrator for it.