About our ​spatial database
We have developed a specialized AI system that simulates thousands of AI-to-AI conversations, leveraging your domain-specific knowledge and requirements. The data from these conversations, along with specific queries, are transformed into vectors within a 1536-dimensional space. Metadata is then associated with these vectors, which can include instructions for executing commands, images, videos, or simple answers.
Subsequently, the data is uploaded to a centralized cloud.When a user generates a query, it is also converted into a vector. The database then retrieves a set number of vectors and their corresponding metadata that are closest to the uploaded vector. Our application extracts the metadata and utilizes it in various ways, such as enabling the Language Model (LLM) to provide more accurate answers or directly providing feedback to the users from the database.
All conversations are stored in the vector database, allowing for detailed vector-based analytics and facilitating shared memory among all devices utilizing our system.
