Amazon has improved its AI video tool, Nova Reel, making it capable of generating videos up to two minutes long.
The upgraded version, Nova Reel 1.1, brings new features aimed at offering more creative control and better consistency across video shots.
Originally launched in December 2024, Nova Reel marked Amazon’s entry into the competitive generative video space, going head-to-head with tools from OpenAI, Google, and others. Now, with version 1.1, users can input prompts of up to 4,000 characters to create multi-shot videos with a smooth and consistent visual style. Each shot lasts about six seconds, making it possible to generate detailed, scene-rich content.
A standout new feature is “Multishot Manual” mode. This allows users to combine a 1280×720 image and a short text prompt (up to 512 characters) to craft up to 20 custom shots, offering a more hands-on way to shape video composition and flow.
Currently, Nova Reel is only available through Amazon’s AWS services, like Bedrock, and access must be requested. However, like many AI video tools, Nova Reel raises ethical questions.
Amazon hasn’t disclosed what video data the model was trained on, nor has it given content creators a way to opt out if their videos were used. Still, the company says it will legally protect AWS users from copyright claims if issues arise from the videos they generate using Nova Reel.
