Yes, Perplexity can make images, and the reason surprises people: it isn't using its own model to do it. The image feature leans on outside generators, which changes what it's good at and where it falls short. Here's how the feature works and when to reach for it.
The short answer
Perplexity generates images, but image creation isn't what the tool is built for. Perplexity is an answer engine first. It searches the web, cites sources, and answers questions with references you can check. Image generation sits on top of that as an extra, not the main event. Wondering whether the feature is even turned on for your account? That depends on your plan.
The feature is usually available on paid plans, and exact availability shifts over time. If you're on the free tier and don't see it, that's why.
So the honest verdict: use Perplexity for images when the image is attached to research or brainstorming. Reach for a dedicated tool when the image itself is the whole job.
How it actually works
When you ask Perplexity for an image, it doesn't run its own image model. It routes the request to an external one, which in recent versions has included models like GPT Image and Nano Banana.
That detail matters because it explains the feature's strengths and limits at once. You get a competent image generator's output, but you get it through Perplexity's simpler interface, without the fine-grained controls a dedicated art tool offers.
You prompt it in plain language, the same way you'd ask a question. There's no separate image console and no parameter panel. Just describe what you want.
Step 1: Ask for an image
- Open Perplexity in a web browser. The web interface has historically carried the feature before mobile.
- Type a request that clearly asks for an image. "Generate an image of..." or "Create a picture of..." works, since the wording signals the intent.
- Wait for the response. Perplexity handles the request and returns a generated visual based on your description.
- Review the result against what you asked for.
If no image comes back, the feature may not be enabled on your account, or your plan doesn't include it. Check your plan before assuming the phrasing is wrong.
Step 2: Write prompts that survive the simpler interface
Because Perplexity gives you fewer controls, your prompt has to do more of the work.
- Describe the subject, the setting, and the style in one clear sentence. "A modern home office with warm natural light, a wooden desk, and a laptop" beats a list of loose keywords.
- Say what the image is for. "A clean blog header for an article about AI search" steers the composition and mood.
- Add a constraint when it matters. If you don't want text baked into the image, say so, because generators still mangle words.
- Refine in follow-up messages, one change at a time, the same way you'd refine a text answer.
The interface rewards clear, plain sentences over clever prompt tricks. That's a real difference from tools built around prompt engineering.
Where it works well
Perplexity's image feature is at its best when the image supports something else you're doing. You research a topic, ask a follow-up question, and then want a quick visual to go with the notes. That flow is smooth because it all happens in one place.
It's also handy for brainstorming. Asking for a rough concept image while you're in the middle of a research thread keeps your momentum, even if the image is only a starting point.
Content planning is another fit. A writer outlining a post can generate a draft header image without leaving the tool, then take it into a proper editor later.
Good for | Not built for |
|---|---|
Quick visuals alongside research | Fine-grained prompt control |
Concept images while brainstorming | Batch generation |
Draft headers for content | Dedicated image API work |
Plain-language requests | Advanced editing and retouching |
Where it falls short
The limits come directly from the design. Because Perplexity isn't a dedicated image studio, you won't find model pickers, seed controls, or the style systems that tools like Midjourney offer. If you need to reproduce an image exactly, or generate fifty variations on a theme, this isn't the tool.
Editing is limited too. Advanced retouching, inpainting, and layered adjustments live in dedicated editors. Perplexity gives you a generated image, not a workspace to refine it in.
And if you're building an app, you'd want an image generation API rather than the built-in feature. Perplexity's version is made for a person asking a question, not a program making thousands of calls.
Putting it to use
Try Perplexity's image feature on a task where you're already researching something, and see whether the visual is good enough to skip a second tool. If it's, you've saved a step. If it isn't, you've learned exactly which jobs belong in a dedicated generator, and that's just as useful.
