AWS offers Amazon Nova, a family of foundation models available through Amazon Bedrock, that includes two creative generation models: Amazon Nova Canvas for images and Amazon Nova Reel for video. Launched in December 2024, these models take text prompts and optional reference images as input and produce studio-quality images or short video clips as output. AWS also offers Nova understanding models that accept text, images, and video as input and generate text, covering tasks such as summarization, document analysis, and conversational AI.

Announced and deployed at AWS re:Invent on December 3, 2024, Amazon Nova is available through Amazon Bedrock after customers request model access. Nova Canvas generates images from text prompts and optional reference images, with controls for color scheme, style, inpainting, and background removal; it is targeted at advertising, marketing, entertainment, and retail use cases. Nova Reel generates short videos from text prompts or reference images, with controls for visual style and camera motion. Both creative models embed invisible watermarks for traceability, and AWS provides uncapped IP indemnity for their outputs. In April 2025, Amazon added Nova Sonic, a speech-to-speech model; in December 2025, Amazon introduced the Nova 2 family with enhanced multimodal reasoning models (Nova 2 Lite and Nova 2 Pro) and Nova 2 Omni, a unified model that accepts text, images, video, and speech as input and generates both text and images.