At BARS Agency, we spend a lot of time inside AI tools. Generating visuals for client campaigns, writing content at scale, prototyping creative concepts — all of it runs through models like Midjourney, FLUX, ChatGPT, and Claude. And the single biggest friction point across all of them has always been the same: getting the prompt right.

Not wrong in a catastrophic way. Just… vague enough to produce generic output. A portrait that looks like every other AI portrait. A landing page copy that reads like a template. A video concept that misses the brief by half a tone.

So we built the tool we kept wishing existed. Today, we’re launching PromptMake — a free AI prompt generator and image-to-prompt tool, built for everyone who works with generative AI daily.

What PromptMake Does

PromptMake solves two related problems in one interface.

Idea to Prompt

Takes a short description — even a single rough sentence — and turns it into a structured, model-ready prompt. You pick your target model (Midjourney, FLUX, ChatGPT, Sora, Claude, Gemini, Runway, and more), select a content type (Image, Video, Text, Audio), and hit Generate. The tool adds the structural vocabulary each model responds to: weighted descriptors for FLUX, camera movement language for Sora, role-task-constraints format for ChatGPT. No guessing, no five rounds of trial and error.

Image to Prompt

Works in reverse. Upload any photo — a reference image, a mood board screenshot, a product shot — and PromptMake analyses its composition, lighting, colour palette, texture, and style, then writes a prompt that would recreate or adapt it in your chosen model. This is particularly useful for brand visual work, where maintaining aesthetic consistency across AI-generated assets is otherwise a manual, painstaking process.

Both tools are free to use without an account. Guests get three generations per day, registered free users get five. Pro unlocks unlimited runs, advanced output formats (JSON, XML, Markdown), and fine-tuning parameters.

Why the Model Layer Matters

One thing we were deliberate about: PromptMake isn’t a generic prompt box. Each model gets its own output format because different models genuinely need different inputs.

Midjourney responds to weighted descriptors and style references. FLUX benefits from precise photographic language. Sora needs camera movement and temporal cues. ChatGPT and Claude work best with clear role definitions, task framing, and format constraints. Writing a FLUX prompt the same way you’d write a ChatGPT prompt is a reliable way to get mediocre output from both.

PromptMake handles this mapping automatically. There are also dedicated landing pages at /for/[model] — so if your workflow is entirely Midjourney, or entirely FLUX, you can bookmark the exact configuration you need and skip the setup every time.

Beyond Prompts: Alt Text and Product Descriptions

Two use cases that sit slightly outside the typical «prompt generation» workflow, but matter a lot in practice:

AI Alt Text Generator — upload any image and get SEO-optimised, accessibility-compliant alt text in seconds. The tool analyses the image content and produces a factual, keyword-aware description that satisfies both WCAG screen reader requirements and Google’s image crawler. For teams managing large content libraries or e-commerce catalogues with hundreds of images, this removes one of the most tedious and commonly skipped steps in the on-page SEO checklist. Missing or thin alt text is still one of the most prevalent technical SEO issues across commercial websites — this makes it trivially easy to fix at scale.

AI Product Description Generator — upload a product photo and get a structured, SEO-ready product description: full listing copy, feature bullet points, or Amazon-style format. The tool extracts visible product features — materials, design, dimensions, category signals — and formats them into copy that serves both conversion and search ranking. Specs not visible in the image can be added via an optional notes field, and multilingual output is supported by specifying the target language in the same field.

Both tools use the same image-to-prompt infrastructure as the core reverse prompt generator, repurposed for specific commercial content needs.

Built on What We Know

This isn’t a side experiment — it’s a direct product of the AI content and SEO work we do at BARS Agency every day. We understand what creative teams and digital agencies actually need from these tools: speed, reliability, output that’s immediately usable rather than requiring a second pass.

PromptMake is live at promptmake.net, no sign-up required to start. Try the text generator, try the image-to-prompt tool, and if it fits your workflow, Pro is $9/month or $79/year.

We’ll be adding models and use cases continuously. If there’s a workflow you’d like to see supported, reach out.