On April 7, 2025, Amazon announced a significant upgrade to its AI-powered video generation tool, Nova Reel.
The latest version, dubbed Nova Reel 1.1, can now produce videos up to two minutes long, marking a big step forward for the tech giant in the increasingly competitive world of generative AI video models.
This development pits Amazon against heavyweights like OpenAI and Google, who are also racing to dominate this cutting-edge space.
First introduced in December 2024, Nova Reel was Amazon’s initial leap into video generation technology. The updated Nova Reel 1.1 brings exciting new capabilities to the table.
According to Elizabeth Fuentes, a developer advocate at Amazon Web Services (AWS), the model can now create “multi-shot” videos with a consistent visual style across different scenes.
This means users can input a detailed prompt of up to 4,000 characters and get a two-minute video made up of six-second shots. Whether you’re crafting a short story, a promotional clip, or an imaginative sequence, this feature offers a lot of creative flexibility.
But that’s not all. Nova Reel 1.1 also introduces a “Multishot Manual” mode, giving users even more control over their projects.
In this mode, you can pair a prompt (up to 512 characters) with a high-resolution image (1280 x 720 pixels) to guide the AI in shaping the video’s composition. The result? Videos with up to 20 distinct shots, tailored more precisely to your vision.
It’s a handy tool for developers or creators who want to fine-tune their output without leaving everything to the AI’s imagination.
Nova Reel 1.1 is exclusively available through AWS platforms, such as Amazon’s AI development suite, Bedrock.
If you’re a developer interested in trying it out, you’ll need to request access, but here’s the good news: AWS automatically approves these requests, so there’s no extra hurdle to jump through.
This clarification came after an earlier report suggested a more restrictive process, which Amazon quickly corrected.
Like many generative AI systems, Nova Reel’s advancements come with some ethical questions. Video-generating models like this one rely on massive datasets of existing videos to “learn” how to create new content.
Some companies have been known to train their models on copyrighted material without permission. When these models accidentally reproduce protected content, it can lead to legal headaches for users.
Amazon hasn’t disclosed the exact sources of Nova Reel’s training data, nor has it offered a clear way for creators to opt out if their work was used.
That said, the company is standing by its customers. AWS has promised to shield users from copyright claims tied to content generated by its models, thanks to its indemnification policy.
It’s a reassuring move, but it doesn’t fully address the broader concerns about transparency in AI development.
Follow TechBSB For More…
