May 1st, 2025 •

Rapidly prototype with gpt-image-1 from Azure OpenAI Service in the Images Playground in Azure AI Foundry

Thasmika Gokal
Thasmika Gokal

Introducing the next-gen Images Playground in Azure AI Foundry

The Images Playground in Azure AI Foundry is a high-fidelity UX sandbox for developers, serving as a reliable testing ground to rapidly prototype with the latest state-of-the-art image generation models – like gpt-image-1 from Azure OpenAI Service – or Stable Diffusion 3.5 Large from Stability AI or Bria 2.3 Fast from Bria AI. Optimize for prompt adherence, latency, and use cases through experimentation; accelerate your “Build-Measure-Learn” development loop; and ship faster with higher confidence.

Ready for enterprise commercial use, images playground in Azure AI Foundry is available for serverless model endpoints from:

Iterate on images, optimize prompts and easily port to code

Whether you’re prototyping new image assets, validating prompts for edge cases, or optimizing output consistency, the images playground removes friction from early experimentation. It’s where ideas are pressure-tested before they are shipped into production. Aligned with production workflows, the images playground is tightly coupled with developer tools like VS Code. Using the “View Code” multi-lingual code samples for your image output, prompts and generation controls within the API structure, what you create in the images playground can be easily ported into VS Code.

  • Iterate faster: Experiment with text and image prompts, adjust generation parameters, and explore editing variations — all in real time with model-specific native API support.
  • Prompt optimization: Debug, tune and re-write prompt syntax with AI (coming soon), compare outcomes across variations you’re testing with, use prebuilt industry prompts (coming soon), and build your own prompt variations available in the playground, grounded in model behavior.
  • Consistent model interface: Everything in images playground mirrors our API structure, so what works here translates directly into code, with predictability and repeatability.

Features

We built images playground for developers who build image generation flows. Images playground is a full featured controlled environment for high-fidelity experiments designed for model-specific APIs – and a great demo interface for your Chief Product Officer and Engineering VP.

  • Model-specific generation controls: Adjust key controls (e.g. number of variants, quality, strength) to deeply understand specific model responsiveness and constraints.
  • Transform with API tooling: Inpainting with text transformation is available for gpt-image-1. Alter parts of your original image with inpainting selection. Use text prompts to specify the change.
  • Port to production with multi-lingual code samples: Use Python, Java, JavaScript, C# code samples with “View Code”. Images playground is your launchpad to development work in VS Code.
  • Side-by-side observations in grid view: Visually observe outputs across prompt tweaks or parameter changes.
  • Azure AI Content Safety integration: With all model endpoints integrated with Azure AI Content Safety, harmful and unsafe images are filtered out prior to being surfaced in images playground.
  • Repeatable: Seed control for specific models enables predictable performance for repeatable experiments with text2img and img2img generation controls. What you build doesn’t stay siloed.

No need to find, build or configure a custom UI to localhost just for image generation, hope that it will automatically work for the next state-of-the-art model, or spend time resolving cascading build errors due to packages or code changes required for new models. The images playground in Azure AI Foundry gives you version-aware access. Build with the latest and most performant models with API updates surfaced in a consistent UI.

How to get started

  1. Sign-in or sign-up to Azure AI Foundry.
  2. Create a Foundry Hub and/or Project.
  3. Create a model deployment for gpt-image-1, Stable Diffusion 3.5 Large or Bria 2.3 Fast from the model catalog or directly from images playground.
  4. Prototype in images playground; iterate over text and image prompts and optimize generation controls for your use case.
  5. Prototype done? Switch to scaled development in VS Code with the Azure OpenAI Service API for gpt-image-1, or the inferencing API for Stability AI and Bria AI deployments.

To Dos

  1. Like and comment with feature requests.
  2. Subscribe to the Foundry Dev Blog.
  3. Post images on X, Discord or LinkedIn.
  4. Scroll for examples of industry use cases and prompts.

Transform enterprise workflows

Start transforming your enterprise workflow with images playground as your immediate next step from model deployment. It is the fastest way to:

  • Validate whether image generation adds value to your use case or product experience.
  • Benchmark quality for domain-specific or brand-aligned use cases with pre-built industry prompt samples.
  • Reduce iteration time before implementation.
  • Collaborate with your team in your Azure AI Foundry Hub/Project to iterate on proof-of-concepts earlier.
Use cases:
  • Pre-production visual ideation: Concept art. Character exploration. Storyboarding. Location scouting. Look-and-feel for make-up, costume previews, lighting and cinematography concepts. Diversity in casting choices.
  • Post-production: CGI character design. VFX planning.
  • Go-to-market: Merchandizing, audience testing and investor pitches, creative experimentation and IP expansion.
Example prompts used above:
  • Left: Create a realistic, pencil-style storyboard sketch of a seasoned female desert nomad running from ancient Egyptian zombies. She faces the camera head on. We see her determined expression.
  • Middle: Design a CGI character for a futuristic sci-fi series: a charismatic, cybernetically enhanced hero in their late 20s, blending human emotion with AI precision. The character should wear a sleek, high-tech armor with glowing energy lines, and have a unique facial feature.
  • Right: Concept art for location scouting of a cliffside monastery carved into sheer rock, glowing with first light through mist-covered mountains. The monastery is stark and minimal—aged stone, weather-beaten prayer flags, and worn footpaths. Vines cover the monastery. A lone figure stands at the edge, robes fluttering, framed by the sun breaking over the horizon. Subtle color grading, fine material detail, and a meditative, cinematic composition.

Gaming

From early prototyping to live operations, image generation empowers gaming studios, developers, and production teams to scale creativity.

Gaming image

Use cases:
  • Production: Concept art. Character exploration. Storyboarding. Pre-visualization and scene composition. 3D model concepts. Character sketches. World-building with realistic textures.
  • Go-to-market: Merchandizing, audience testing and investor pitches, creative experimentation and IP expansion.
Example prompts used above:
  • Left: Create a realistic gaming character design colored sketch of a fantasy character. Storyboard
  • Middle: Create a realistic gaming fantasy character design in front of an exploding volcano. We see a close-up of her face with determined emotion. A red light flashes on her face.
  • Right: Minecraft scene with character in underground cave with glowing redstone tech, waterfalls and minecart tracks.

Retail & Advertising

Enable personalized, high-quality visuals that can improve customer engagement and streamline creative workflows that enable consumers with a more immersive and intuitive experience.

 

Use cases:
  • Product visualization through high quality online listing.
  • Ads and promotional content.
  • Customization options with visual representations of customized products.
  • Virtual try-ons. Visual campaigns. Digital ads. Social media content.
Example prompts used above:
  • Left: Male models advertising the latest luxury watch. NYC is in the background.
  • Middle: Social media advertisement for luxury perfume.
  • Right: Advertising luxury Indian jewelry.

Real estate, architecture & urban planning

Enhance visualization for virtual staging, marketing, visual proposals and realistic renderings for real estate and sustainable urban planning for commercial, communities and cities.

Image Real estate

Use cases:
  • Property visualization. Virtual tours. Design realistic renders of homes or commercial buildings. Visualize renovations or improvements. Digitally stage.
  • Experiment with materials, lighting and designs to meet sustainability development goals.
  • Simulate city developments, assess sustainable impacts. Visualize cityscapes, parks, transportation systems and public spaces.
Example prompts used above:
  • Left: Interior of a Spanish home on the Malibu coastline.
  • Middle: Interior of a sustainable office building featuring exposed timber beams, passive lighting, recycled materials, and greenery. Realistic materials, ambient lighting, and human-scaled design.
  • Right: Urban planning for Brisbane, Australia in 2050.

Industrial development & manufacturing

30 04 2025 13 24 42 REC image

Use cases:
  • Rapid Prototyping Visuals: Generate early visualizations of product concepts before CAD modeling.
  • Alternative Design Explorations: Create multiple design variations quickly to explore form, texture, and color.
  • Industrial Aesthetics Testing: Visualize a product in various material finishes, such as brushed steel, matte plastic, or anodized aluminum.
  • Factory Layout Visualization: Generate mockups of factory floor configurations and assembly line layouts.
  • Tooling & Fixture Concepts: Visualize jigs, fixtures, or tooling setups before detailed engineering.
  • Custom Part Mockups: Generate visuals for niche, low-volume components to test with vendors or clients.
  • Packaging Prototypes: Quickly visualize how a product would look in different packaging designs.
  • Regulatory/Labeling Compliance: Mockup variations of safety labels or regulatory markings for different markets.
  • Photorealistic Renders: Generate images for use in brochures, websites, and ads without the cost of physical prototypes.
  • Customization Demos: Show clients or customers what a customized version of a product could look like.
  • Operator Training: Visualize equipment or machinery for simulation-based training modules.
  • Maintenance Guides: Create visuals for user manuals or AR/VR-based repair instructions.
Example prompts used above:
  • Left: Quantum computing chip.
  • Middle: 1st person view of AR/VR in a manufacturing plant.
  • Right: Wide-angle interior view of a modern smart factory with automated assembly lines, robotic arms, IoT-connected machines, and real-time dashboards on mounted screens. Clean layout, realistic lighting, and industrial textures.

Education

Use cases:
  • Virtual learning environments: Generate educational materials. Interactive diagrams and models. Visual simulations for complex subjects. Scenario-based training preparing people for high-pressure environments. Customizable visuals can cater to diverse learning needs and provide personalized content.
Example prompts used above:
  • Left: High-resolution educational illustration of the human brain with labeled lobes, nerves, and functions. Clean white background, medically accurate detailing, color-coded for easy comprehension.
  • Middle: Reconstruction of a bustling Roman marketplace with merchants, citizens in togas, stone buildings, and carts of produce. Realistic textures, historically accurate architecture and clothing, warm afternoon lighting.
  • Right: Visual representation of quantum unit circle.

What’s next

We are continuing to invest in capabilities that matter to developers – all grounded in real model behavior and consistent API support. We will update you in this blog’s changelog to keep you updated on the latest action.

Azure AI Foundry Images Playground is where your image generation prototype starts. Whether you are building for automation, tooling, or AI-native applications, Images Playground in Azure AI Foundry is where you can iterate on images, optimize prompts and port generations into code.

Create the Future of AI

Category
Azure AI Foundry

0 comments

Read next