CPU3DAI tools for 3D, video and audio

Sora: video generator — what it does, formats, and rights to the result

Sora is a cloud-based video generator from OpenAI that turns text, images, and existing video clips into new scenes. The service is aimed at producing both realistic and stylized animated content with special effects.

What it does

The Sora 2 model accepts three types of input: a text description, a static image, or an existing video. Based on these instructions, it generates a clip. As described by the author, the model can create a wide range of content, from realistic scenes to animation and visual effects.

Users can choose from six available resolutions: 480x480, 720x720, 1080x1080, 1280x720, 1920x1080, and 720x1280 pixels. The duration of the final clip is limited to 4, 8, or 12 seconds. The finished result is delivered in MP4 format.

How to use it

Access to the generator is provided in two ways: through the web interface on the Microsoft Foundry portal and through an application programming interface (API). The developer notes that Sora 2 uses the Azure OpenAI v1 API scheme, compatible with the model's native scheme.

The service operates exclusively as a cloud solution. The source code is not published, and trained model files or weights are not distributed. Sora cannot be run locally.

Rights to the output

The authors do not specify the terms of rights transfer for generated videos in the technical documentation. This aspect is governed by the service's user agreement, which should be reviewed separately.

Who it suits

The tool will be useful for those who need short video drafts for concept art, storyboards, social media, or advertising creatives. Support for image-to-video and video-to-video modes allows it to be used for animating static content or processing existing material.

The service is not suitable for those looking for a free or local solution without cloud dependency, nor for those who need long clips exceeding 12 seconds. The lack of access to code and weights makes it unsuitable for research tasks involving fine-tuning or architecture modification.

The platform offers a fixed set of resolutions and durations, which simplifies predicting the result but imposes limitations on non-standard formats. Before starting, it makes sense to check the current terms of use on the developer's website.

Sora 2 pipeline Video generation from text, image, or video Input Text, image or video text-to-video image-to-video Parameters Resolution Duration 480×480 … 1920×1080 4 / 8 / 12 s Sora 2 model OpenAI cloud service Azure OpenAI API 5 API endpoints video-to-video realistic scenes Output Finished video animation, effects MP4 export Interaction methods API (Azure OpenAI v1) • Browser (Microsoft Foundry portal) Sora 2 — OpenAI — cloud service, closed source Docs: learn.microsoft.com — Azure AI Foundry
How the Sora pipeline works. The diagram is drawn from the tool’s fact sheet.

Fact sheet

Tasktext-to-video, image-to-video, video-to-video per the website source
How to useAPI, browser (Microsoft Foundry portal) per the website source
Inputtext, image, video per the website source
Resolution480×480, 720×720, 1080×1080, 1280×720, 1920×1080, 720×1280 per the website source
Clip length4, 8, 12 с per the website source
Export formatsMP4 per the website source
API documentationlearn.microsoft.com/en-us/azure/ai-foundry/openai/concepts/video-generation per the website source
Changes often — as of 2026-08-05
Model versionSora 2 per the website source
LimitsVideo generation typically takes 1 to 5 minutes per the website source

Values are collected automatically from official sources and were checked on 2026-08-05. Each one links to its source, and values taken from the developer’s pages also carry a verbatim quote — hover over the note. Pricing and versions are shown as of the check date and change most often; verify on the vendor’s site before buying.

See also