# Stable Diffusion AI online

> stablediffusion.fr is a free web platform for AI image generation and open AI models: a public Stable Diffusion
> inpainting tool and an AI upscaler running on the site's own GPU, text-to-image demos (FLUX 3 Klein, FLUX.2,
> Qwen-Image 2.1, Stable Diffusion 3, SDXL, SD 1.5), image-to-video demos (Wan 2.2, MiniMax), chat interfaces (GPT-4
> family via the OpenAI API, Qwen, Llama 3, Mixtral, DeepSeek-R1 in the browser), and reference material for prompt
> engineering: 1,731 artist styles compared with identical seeds, 3,903 artist names with samples, 94 art styles
> compared with FLUX.2, about 2,900 prompts with their images across the galleries and the style page, and 132 fully
> detailed prompts with all their generation parameters. No account is required
> for public tools; a free account unlocks the full txt2img interfaces. Home page and GPT-4 page in 22 languages,
> inpainting in 6. Maintained by Anthony Clicheroux.
> Contact: contact@stablediffusion.fr — https://x.com/tony33680 — Last updated: 2026-10-20

## Image tools (run on the site's GPU, no account)
- [Stable Diffusion web UI](https://stablediffusion.fr/webui): Type a prompt and get a Stable Diffusion 1.5 image without an account (20 steps, Euler a); then change the format (512×512, 512×768, 768×512), the CFG scale and the seed. One image every 20 seconds, 30 per day.
- [Inpainting](https://stablediffusion.fr/inpainting): Upload an image, paint a mask, describe the result; Stable Diffusion regenerates only the masked area. Images up to 1024 px per side, 1 to 30 steps.
- [Upscaling](https://stablediffusion.fr/upscaling): Enlarge an image up to 2x with ESRGAN, R-ESRGAN 4x+, R-ESRGAN Anime6B, SwinIR, DAT, ScuNET or Lanczos.
- [Stable Diffusion txt2img (members)](https://stablediffusion.fr/txt2img): Full text-to-image with Stable Diffusion 1.5, negative prompt, seed, size. Free account required ([sign up](https://stablediffusion.fr/signup), [log in](https://stablediffusion.fr/login)).
- [Stable Diffusion XL txt2img (members)](https://stablediffusion.fr/txt2img-xl): SDXL 1.0 text-to-image at 1024×1024, free account required.

## Image demos (embedded open-source demos)
- [FLUX 3 Image Klein](https://stablediffusion.fr/flux3): Text-to-image demo of Black Forest Labs' open FLUX 3 Image Klein (5.6B, native 4K, readable text), free, no account.
- [Qwen-Image 2.1](https://stablediffusion.fr/qwen-image): Text-to-image and image editing demo of Alibaba's open Qwen-Image 2.1, legible text in any language.
- [What is Qwen-Image?](https://stablediffusion.fr/guide/what-is-qwen-image): Alibaba's open-weights image model: what Qwen-Image 2.1 can do (2K, editing with 10 references, RGBA, text), how it works compared with Stable Diffusion (Qwen3-VL encoder, single-stream transformer, 64-channel latent, flow matching), the versions and the licence.
- [Qwen-Image 2.1 guide](https://stablediffusion.fr/guide/qwen-image-2-1): how to prompt it, the settings that work (CFG, steps, APG and FreSca), native 2K and generation times, see the Guides section.
- [FLUX.2](https://stablediffusion.fr/flux): Text-to-image demo of FLUX.2 [dev] by Black Forest Labs, sharp detail and readable text.
- [Stable Diffusion 3 medium](https://stablediffusion.fr/demo): Text-to-image demo of Stable Diffusion 3 Medium by Stability AI.
- [Stable Diffusion XL 1.0](https://stablediffusion.fr/xl): SDXL 1.0 text-to-image demo, 1024×1024 images.
- [Stable Diffusion 3](https://stablediffusion.fr/3): Presentation of Stable Diffusion 3 and its improvements (prompt understanding, text rendering).
- [SDXL Turbo image-to-image](https://stablediffusion.fr/sdxl-turbo): Real-time image-to-image with the distilled SDXL Turbo model.
- [Open DALL-E](https://stablediffusion.fr/dalle): Open-source DALL-E style image generation.
- [AI QR code generator](https://stablediffusion.fr/qrcode): Scannable QR codes blended into AI images with Stable Diffusion ControlNet.
- [Prompt generator](https://stablediffusion.fr/prompt-generator): Turn a few words into a detailed Stable Diffusion prompt (MagicPrompt).

## Video, animation and audio demos
- [MiniMax-H3](https://stablediffusion.fr/minimax): Short video with a synchronized soundtrack from a text prompt or from a first and last frame, MiniMax's open video model.
- [Wan 2.2 image to video](https://stablediffusion.fr/wan): Turn a photo into a 5-second clip with Alibaba's open Wan 2.2 model.
- [Wan 2.2 guides](https://stablediffusion.fr/guide/wan-image-to-video): image to video and [text to video](https://stablediffusion.fr/guide/wan-text-to-video), see the Guides section.
- [Text to video](https://stablediffusion.fr/txt2vid): Short clips from a text prompt (LaVie).
- [Image to video](https://stablediffusion.fr/img2vid): Animate a still image (AnimateLCM).
- [Real-time latent consistency](https://stablediffusion.fr/video): Real-time image generation with a Latent Consistency Model.
- [LivePortrait](https://stablediffusion.fr/live-portrait): Animate a portrait photo from a driving video.
- [MusicGen](https://stablediffusion.fr/musicgen): Music generation from a text description (Meta MusicGen).

## Chat and language models
- [GPT-4](https://stablediffusion.fr/chatgpt4): Free chat interface to OpenAI's GPT-4 family through the OpenAI API (currently the GPT-4o mini model), streaming answers, no account. Available in 22 languages at /<lang>/chatgpt4.
- [GPT-3.5](https://stablediffusion.fr/chatgpt3): Lighter, faster chat interface on the same API.
- [GPT-4 turbo](https://stablediffusion.fr/gpt4turbo) and [GPT-5](https://stablediffusion.fr/gpt5): Additional chat pages.
- [Chat with Jack](https://stablediffusion.fr/chatjack): The same assistant playing Captain Jack Sparrow.
- [Llama 3 8B](https://stablediffusion.fr/llama3): Chat with Meta's open-weights Llama 3 8B.
- [Llama 2 70B](https://stablediffusion.fr/llama2): Chat with Meta's Llama 2 70B.
- [Mixtral 8x7B](https://stablediffusion.fr/mixtral): Mistral AI's sparse mixture-of-experts model.
- [Qwen 3.8 chat](https://stablediffusion.fr/qwen): Free chat with Alibaba's Qwen 3.8 (27B) for questions, code and reasoning, no account.
- [Qwen3-VL](https://stablediffusion.fr/qwen3): Ask questions about an image with Alibaba's open vision-language model: description, text reading, charts, chat.
- [DeepSeek-R1](https://stablediffusion.fr/deepseek-r1): Reasoning model running entirely in the browser (Transformers.js, ONNX Runtime Web); nothing is sent to a server.
- [Zephyr 7B](https://stablediffusion.fr/zephyr), [StableLM](https://stablediffusion.fr/stablelm), [StableVicuna](https://stablediffusion.fr/stablevicuna), [Chatbot Arena](https://stablediffusion.fr/vicuna), [Stable Code Instruct 3B](https://stablediffusion.fr/stablecode), [Open GPT](https://stablediffusion.fr/opengpt): Other open-model chat demos.
- [GPT detector](https://stablediffusion.fr/chatgpt-detector): Estimate whether a text was written by an AI model.

## Prompt engineering references
- [FLUX.2 styles](https://stablediffusion.fr/flux2-styles): art styles rendered with FLUX.2 [klein] on the same six scenes and the same seed, one row per style, scenes selectable, style sentence copyable.
- [Artist styles for Stable Diffusion XL](https://stablediffusion.fr/artists): 1,731 artists, each with 4 images generated from the same seed and prompt, only the artist name changing; 10 alphabetical pages (https://stablediffusion.fr/gallery/artists-page02 … page10), search and one-click copy of "by <artist>".
- [Artist-style comparison](https://stablediffusion.fr/artist-style): 3,903 artist names with 4 samples each, 100 per page (?page=N), searchable.
- [Prompt examples](https://stablediffusion.fr/prompts): 132 fully detailed prompts with all their generation parameters (model, sampler, steps, CFG scale, seed, size, negative prompt), grouped by theme, filterable by model (Stable Diffusion 1.5, SDXL, FLUX.2, Qwen-Image 2.1), copyable.
- [Image gallery](https://stablediffusion.fr/images): Every image of the site with its prompt, newest first, about 2,300 (1,000 in the themed galleries, 1,100 shared by the community in 2022 and 2023), 48 per page, full-text search on prompts. With the FLUX.2 style page, about 2,900 prompts with their images on the site.
- [Galleries by theme](https://stablediffusion.fr/gallery): art, architecture, animals, anime, crafts and illustration, fantasy, fun, interior design, landscape, misc, sci-fi, animated; community galleries by week (2022) and by month (Nov 2022 – May 2023), for example https://stablediffusion.fr/gallery/landscape or https://stablediffusion.fr/gallery/2023-05.

## Guides
Eleven short, practical guides in 21 languages (English at /guide/<slug>, others at /<lang>/guide/<slug>; index: https://stablediffusion.fr/guide).
- [What is Stable Diffusion?](https://stablediffusion.fr/guide/what-is-stable-diffusion): How it works step by step, with real renders: diffusion, the latent space and the VAE, CLIP and cross-attention, guidance scale, negative prompt and seed, the versions (1.5, SDXL, 3, 3.5) and how to use it free.
- [Prompt guide](https://stablediffusion.fr/guide/prompt): Subject, style, light and quality in the right order, an annotated example, weights, mistakes to avoid.
- [Negative prompt](https://stablediffusion.fr/guide/negative-prompt): What to leave out, ready-to-use lists, weights, and when it does nothing.
- [Seed](https://stablediffusion.fr/guide/seed): Reproduce, compare and vary images with the same seed and settings.
- [LoRA](https://stablediffusion.fr/guide/lora): Small files that add a style, a character or an object; how to use, weight and combine them.
- [Install Stable Diffusion locally](https://stablediffusion.fr/guide/install): Hardware, Python and Git, Automatic1111, model download, first launch, common errors.
- [SDXL vs Stable Diffusion 3 vs SD 1.5](https://stablediffusion.fr/guide/sdxl-vs-sd3): Resolution, VRAM, prompt understanding, text rendering, speed, licence, which one to pick.
- [Wan 2.2 image to video](https://stablediffusion.fr/guide/wan-image-to-video) and [Wan 2.2 text to video](https://stablediffusion.fr/guide/wan-text-to-video): motion prompt, settings, limits, local run.
- [What is Qwen-Image?](https://stablediffusion.fr/guide/what-is-qwen-image): what Qwen-Image 2.1 can do, how it works compared with Stable Diffusion, versions and licence.
- [Qwen-Image 2.1](https://stablediffusion.fr/guide/qwen-image-2-1): prompt structure, measured settings, resolution and steps trade-offs, what slows it down, local run in ComfyUI, the guided Turbo ComfyUI workflow to download.

## Languages
- Home page: [English](https://stablediffusion.fr/), [Français](https://stablediffusion.fr/fr), [Español](https://stablediffusion.fr/es), [Deutsch](https://stablediffusion.fr/de), [Italiano](https://stablediffusion.fr/it), [Português](https://stablediffusion.fr/pt), [Nederlands](https://stablediffusion.fr/nl), [Polski](https://stablediffusion.fr/pl), [Svenska](https://stablediffusion.fr/sv), [Dansk](https://stablediffusion.fr/da), [Magyar](https://stablediffusion.fr/hu), [Ελληνικά](https://stablediffusion.fr/el), [Türkçe](https://stablediffusion.fr/tr), [Українська](https://stablediffusion.fr/uk), [العربية](https://stablediffusion.fr/ar), [עברית](https://stablediffusion.fr/he), [日本語](https://stablediffusion.fr/ja), [한국어](https://stablediffusion.fr/ko), [中文](https://stablediffusion.fr/zh), [ไทย](https://stablediffusion.fr/th), [Tiếng Việt](https://stablediffusion.fr/vi), [অসমীয়া](https://stablediffusion.fr/as)
- GPT-4 page: same 21 language codes as prefixes, e.g. https://stablediffusion.fr/de/chatgpt4, https://stablediffusion.fr/ja/chatgpt4. Inpainting: https://stablediffusion.fr/fr/inpainting, /es/, /de/, /ja/, /zh/.

## About and policies
- [About](https://stablediffusion.fr/about): The platform and its maintainer.
- [Privacy policy](https://stablediffusion.fr/private-policy): Data collected by the site and its partners, cookies and consent.
- Adult content and violence are not allowed anywhere on the site: such prompts and images are refused on every tool, public or members-only.
- Public generation tools share one GPU: jobs are queued, progress is streamed, one submission per IP every 20 seconds, results are kept a few minutes. Generated images belong to their authors; prompts are shown next to community images so they can be reused.

## Optional
- [llms-full.txt](https://stablediffusion.fr/llms-full.txt): Detailed description of every section, for LLM agents.
- [Sitemap](https://stablediffusion.fr/sitemap.xml): All indexable URLs.
- [Image sitemap](https://stablediffusion.fr/sitemap-images.xml): Full-size gallery and artist-style images, each attached to the page that displays it.
- [robots.txt](https://stablediffusion.fr/robots.txt): Crawling rules (everything allowed, search, AI answers and AI training permitted).