In this example here-above, Lightkey is able to identify ‘SpaceX’ as a word indicating that the context is aerospace engineering, ‘launch’ as the action taken, and ‘unmanned’ as a word relevant to aerial objects. Therefore, the relevant prediction would be ‘spacecraft’. The same analysis could be further applied to the next predictions such as “to” and “Mars”. This is just one example, but in practice Lightkey’s AI technology can identify meaningful relationships in many areas (or content domains).​​​

​

Please note that Lightkey is not a generative AI solution. Unlike tools designed to replace the content creation process, Lightkey’s goal is to enhance and accelerate writing through personalized, user-driven predictions. Once installed, Lightkey’s suggestions are intentionally minimal and non-disruptive, focusing first on learning each user’s unique writing style. Over time, as it becomes more familiar with the user’s tone, vocabulary, and context, Lightkey gradually increases the number of its suggestions and spelling corrections.

​

Behind the scenes, Lightkey operates using seven distinct language models. The core AI language model is static, meaning it provides a foundational understanding but does not learn or adapt. The other models are dynamic as they continuously evolve based on the user's typing patterns, allowing Lightkey to personalize predictions with increasing precision.​

​

Importantly, Lightkey is an offline solution, which means that all processing occurs locally on the user’s device, ensuring that work and data remain completely private and secure at all times.