Video is the king now in the world of content creation. Creating high-quality videos usually requires significant resources such as time, money and manpower. Generative AI is easing the video making process. It is promising to revolutionize video production by automating the process as well as cutting the costs drastically. However, even the most advanced models such as OpenAI’s Sora is often falling short and videos produced are either inaccurate or unfit for professional due to quality concerns and copyright issues.
Bengaluru-based startup Unscript has emerged in the scene. It is making waves by offering a video generation platform that competes with OpenAI and other AI giants. The platform can transform a single photo into a lifelike video within two minutes. It is believed to be a new standard in AI-driven video creation.
The Unscript model stands out as it is mainly focusing on human-like video content. OpenAI’s Sora often generates abstract or fantastical videos like dolphins cycling in the ocean. Unscript’s AI is trained for realistic human interactions.
Unscript co-founder Ritwika Chowdhury describes the platform as a “Canva for videos.” It offers an end-to-end solution from video shooting to final production. It has the ability to automate everything in the video creation process. It has turned incredibly attractive for businesses which are looking to save time and reduce costs.
The Unscript edge comes from its robust training data and model architecture. The startup has developed an encoder-decoder plus diffusion-based model. Its system has been trained on a substantial dataset of more than 1,000 hours. Its extensive training allows the platform to generate videos with precise lip-syncing, eye movements, facial expressions and body language. All these are tailored to the individual being portrayed.
