Persona is a user context engine, that maps human behavior for next gen personalization. It takes in a user’s digital footprint and builds a knowledge graph that dynamically models the user’s mindspace and personality.

The personal knowledge graph construction takes inspiration from computational theory of mind, associating the narratives, memetics, symbols that form the user’s identity and personality.

Why knowledge graphs with vector databases? Human memory is associative, and self organizing. And graph is the closest data structure to imitate this behavior. Memory is also semantic, and thus needs vector representations.

The key insight behind Persona is that with LLMs, it’s possible to learn latent cognitive patterns about the user’s behavior, store them in a structured way, and provide rich and dynamic user context to LLMs and other AI agents for more personalized interactions.

Persona is designed for LLM driven consumer interfaces — conversational or generative — to deliver adaptive personalization to their users. But it can be extended to traditional app interfaces as well.

Quickstart

Get started with Persona. Setup your deployment and start indexing your data.

Intelligent Identities

Read a blog on the design of Persona, and the need for an identity based on language and memetic theory.