Post the image together with the recipe behind it. Follow creators you trust, then open their workflows as ready-to-run tabs and make the next version your own.
Browse and remix public work without an account. Sign in only to save, follow, or publish.

strxet.nyc
@strxet_nyc · 2h

nyxxa.art
@nyxxaart · Jul 10

ember.lens
@ember_lens · 7h
A community built around the image
See the artwork first, then open the prompt, model notes, and reusable workflow behind it.
When a creator shares a workflow, you do not have to reverse-engineer the image. Open their setup in your Builder, adjust it locally, and keep the original creator connected to your remix.

POSITIVE PROMPT
masterpiece, cinematic portrait, golden hour, warm rim light, 35mm film grain
Prompt Builder is the local-first workspace where you turn experiments into repeatable systems. Compose prompts from reusable blocks, recover past settings, refine the result, then share only when you choose.
Hit the ✦ button to enhance a rough prompt or describe an idea in plain English and get a clean, structured prompt back — ready to tweak in the editor.
Style-by-style guides with copy-ready example prompts — open any one straight in the builder, then tweak, weight and generate.
🚀 New to AI image generation? Start with the beginner's guide →
A believable AI portrait is mostly camera language and a ruthless negative prompt. Here is the exact structure — lens, light, skin and eyes — plus prompts you can open in the builder and run.
Appetising food images come from styling and light, not adjectives. Here is how to prompt steam, freshness and that editorial-menu look — with prompts you can run in one click.
Anime models speak danbooru tags, not sentences. Here is the tag order that works on Pony and Illustrious, the quality blocks each one wants, and prompts you can open and run.
A cinematic image is a series of deliberate choices — shot size, lens, lighting ratio, color grade. Borrow the cinematographer’s vocabulary and your renders stop looking like stock art.
Diffusion models do not make production logos, but they are great at concepts and marks. The trick is forcing simplicity — flat, minimal, vector-style — instead of the busy default.
Epic landscapes are about scale, atmosphere and a single focal point. Here is the matte-painting structure concept artists use — plus prompts you can open and run.
Commercial product shots are about clean light, a deliberate surface and a background you can cut out. Here is the studio structure — plus prompts you can open and run.
Strong character concepts read from the silhouette in and specify materials, not just "cool armor". Here is the designer’s structure — plus prompts you can open and run.
Convincing architecture renders come from naming the style, the materials and the light — not from "modern house". Here is the archviz structure, with prompts you can run.
Designed for power users who build prompts all day.
A built-in social network for AI image creators: post prompts & AI art, follow creators, like and reshare, and install anyone's full workflow as a ready-to-run tab.
Enhance any prompt, or turn a plain-English idea into a full, structured prompt — powered by your own OpenAI, Anthropic or OpenRouter key, or a 100% local Ollama / LM Studio model. Your keys, your models, your privacy.
Buttons insert reusable text snippets at your cursor — no fragile toggle state, no sync bugs. You stay in full control of every token.
Your color-coded prompt library on the left, a live editable output on the right. Click to build, type to refine.
Buttons appear only when relevant to your selected Model or LoRA. Fuzzy search the entire library instantly.
Highlight any token and press Ctrl/Cmd + ↑/↓ to wrap it in attention weights like (token:1.1) — no manual typing.
Start in seconds with curated libraries for Pony, Illustrious, SDXL, Flux and realistic photo — or build your own from scratch.
Live CLIP token counter, duplicate & conflict detection, and an offline booru tag autocomplete as you type.
Use {red|blue|green} variants and __named__ wildcards, then roll a fresh concrete prompt with one click.
Drag in any image generated by A1111 or ComfyUI and instantly recover its positive & negative prompt.
Connect a local A1111/Forge or ComfyUI server and generate images right from the prompt — no copy-pasting into another window.
Run parallel prompt workflows in separate tabs — each with its own model, LoRA and prompt state — and group them into collapsible folders.
Your library, tabs and folders follow you seamlessly between the web app and desktop. Set it up once, use it everywhere.
Start free. Upgrade only when you need more.
Billed in EUR · cancel anytime
Have a promo code? Open the app → account menu → Redeem a code.
Native window, offline access, same great features. Or just use it in your browser.
Yes. The full prompt-building engine is free forever on web and desktop. Pro unlocks unlimited tabs, cloud sync, and early access to new features.
Browse creators, collect workflows, then open the builder when you want to remix.
Share more than the final image
PromptHub keeps the art and the process together. Publish a prompt, attach a reusable workflow, credit the tools and model, and give other creators something they can actually learn from.
Post images with the positive and negative prompts that made them
Share complete workflows and button libraries people can reuse
Follow creators, save discoveries, like, and reshare
Discuss techniques and get help in the creator forum