Introducing Sora and video playground in Azure AI Foundry
The video playground in Azure AI Foundry is your high-fidelity testbed for prototyping with cutting-edge video era fashions – like Sora from Azure AI Foundry Fashions – prepared for business use. Learn our Tech Group launch weblog on gpt-image-1 and Sora.
Fashionable improvement entails working throughout a number of programs—APIs, providers, SDKs, and information fashions—usually earlier than you’re prepared to completely decide to a framework, write exams, or spin up infrastructure. Because the complexity of software program ecosystems will increase, the necessity for protected, light-weight environments to validate concepts turns into crucial. Video playground was constructed to satisfy this want.
Goal-built for builders, video playground affords a managed atmosphere to experiment with immediate buildings, consider mannequin consistency relative to immediate adherence, and optimize outputs for trade use circumstances. Whether or not you’re constructing AI-native video merchandise, instruments, or reworking your enterprise workflows, video playground enhances your planning and experimentation — so you possibly can iterate quicker, de-risk your workflows, and ship with confidence.
Quickly prototype from immediate to playback to code
Video playground affords an on-demand, low-friction-setup atmosphere designed for speedy prototyping, API exploration and technical validation with video era fashions. Consider video playground as your high-fidelity prototyping atmosphere – constructed that can assist you construct higher, quicker and smarter – with no configuration of localhost, importing clashing dependencies or worrying about compatibility between construct and mannequin.
Sora from Azure OpenAI is the primary launch for video playground – with the mannequin coming with its personal API – a novel providing obtainable for Azure AI Foundry customers. Utilizing the API in VS Code permits for scaled improvement in your VS Code atmosphere on your use case as soon as your preliminary experimentation is completed within the video playground.
- Iterate quicker: Experiment with textual content prompts and modify era controls like side ratio, decision and length.
- Immediate optimization: Debug, tune and re-write immediate syntax with AI, visually examine outcomes throughout variations you’re testing with, use prebuilt trade prompts, and construct your individual immediate variations obtainable within the playground, grounded in mannequin habits.
- Constant interface for API: All the pieces in video playground mirrors the mannequin API construction, so what works right here interprets straight into code, with predictability and repeatability.
Options
We constructed video playground for builders who need to experiment with video era. Video playground is a full featured managed atmosphere for high-fidelity experiments designed for model-specific APIs – and a fantastic demo interface on your Chief Product Officer and Engineering VP.
- Mannequin-specific era controls: Modify key controls (e.g. side ratio, length, decision) to deeply perceive particular mannequin responsiveness and constraints.
- Pre-built prompts: Get impressed on how you should use video era fashions like Sora on your use case. Within the pre-built prompts tab, there’s a set of 9 curated movies by Microsoft.
- Port to manufacturing with multi-lingual code samples: Within the case of Sora from Azure OpenAI – this displays the Sora API – a novel providing obtainable to Azure AI Foundry customers. Utilizing the “View Code” multi-lingual code samples (Python, JavaScript, GO, cURL) on your video output, prompts and era controls that mirror the API construction. What you create within the video playground could be simply ported into VS Code as a way to proceed scaled improvement in VS Code with the API.
- Aspect-by-side observations in grid view: Visually observe outputs throughout immediate tweaks or parameter modifications.
- Azure AI Content material Security integration: With all mannequin endpoints built-in with Azure AI Content material Security, dangerous and unsafe movies are filtered.
See a demo of those options and Sora in video playground in our devoted breakout session at Microsoft Construct 2025 right here.
No want to seek out, construct or configure a customized UI to localhost for video era, hope that it’s going to mechanically work for the following state-of-the-art mannequin, or spend time resolving cascading construct errors resulting from packages or code modifications required for brand spanking new fashions. The video playground in Azure AI Foundry provides you version-aware entry. Construct with the newest fashions with API updates surfaced in a constant UI.
What to check for in video playground
When utilizing video playground, as you propose your manufacturing workload, contemplate the next as you’re visually assessing your generations:
- Immediate-to-Movement Translation
- Does the video mannequin interpret my immediate in a approach that makes logical and temporal sense?
- Is movement coherent with the described motion or scene? How might I take advantage of Re-write with AI to enhance my immediate?
- Body Consistency
- Do characters, objects, and types stay constant throughout frames?
- Are there visible artifacts, jitter, or unnatural transitions?
- Scene Management
- How nicely can I management scene composition, topic habits, or digital camera angles?
- Can I information scene transitions or background environments?
- Size and Timing
- How do totally different immediate buildings have an effect on video size and pacing?
- Does the video really feel too quick, too sluggish, or too brief?
- Multimodal Enter Integration
- What occurs after I present a reference picture, pose information, or audio enter?
- Can I generate video with lip-sync to a given voiceover?
- Put up-Processing Wants
- What degree of uncooked constancy can I count on earlier than I want modifying instruments?
- Do I have to upscale, stabilize, or retouch the video earlier than utilizing it in manufacturing?
- Latency & Efficiency
- How lengthy does it take to generate video for various immediate varieties or resolutions?
- What’s the cost-performance tradeoff of producing 5s vs. 15s clips?
Run Sora and different fashions at scale utilizing Azure AI Foundry—no infrastructure wanted. Study extra in our current Microsoft Mechanics video that shares extra concerning the Sora API in motion:
Get began now
- Signal-in or sign-up to Azure AI Foundry.
- Create a Foundry Hub and/or Mission.
- Create a mannequin deployment for Azure OpenAI Sora from the Foundry Mannequin Catalog or straight from video playground.
- Prototype in video playground; iterate over textual content prompts and optimize era controls on your use case.
- Prototype performed? Change to scaled improvement in VS Code with the Sora from Azure OpenAI API.
Create with Azure AI Foundry
- Learn our Tech Group weblog on gpt-image-1 and Sora.
- Get began with , and soar straight into
- Obtain the
- Take the
- Evaluation the
- Preserve the dialog getting in and
