GetProfile is a self-hosted and open-source solution that provides user profiles and long-term memory for AI agents. Unlike generic memory solutions that store blobs of text, GetProfile extracts natural language summary, typed traits with confidence scores and relevant memories with importance levels into structured user profiles stored in your own PostgreSQL database.

The product functions as a drop-in LLM proxy that gives an AI model persistent memory and structured user understanding. It automatically adds a system message with user profile summary, traits, and relevant memories to each prompt. GetProfile offers fully customizable user traits schema for extraction and injection, allowing you to define exactly what user traits matter for your application.

The system works by having your agent send a request to GetProfile proxy, which then enriches the request with user profile and memory. The enriched request is forwarded to the LLM provider, which generates a response sent back to your agent while GetProfile updates the user profile and memory in the background. This approach provides minimal latency impact and seamless integration with existing workflows.

GetProfile enhances LLM-powered applications with personalized and context-aware interactions without requiring changes to your existing OpenAI-based code. It eliminates overloaded prompts and context windows, long history of messages, and blackbox solutions with unpredictable behavior by providing relevant, structured information you define.

The solution is designed for developers building AI applications who need persistent user memory and structured profiles. It's built with Hono for the lightweight proxy and uses PostgreSQL for persistent profiles storage, featuring minimum external dependencies and easy scaling with Docker Compose.