AI Memory
Provide persistent, ranked, and cryptographically attested memory storage across diverse AI models and agent workflows.
AI Memory is a universal, persistent memory substrate built in a single Rust binary designed to seamlessly integrate with any MCP-capable client, including Claude, ChatGPT, Grok, and local models. It enables AI assistants to retain architecture decisions, user preferences, and learned corrections across sessions using either local SQLite or scalable PostgreSQL with Apache AGE and pgvector. Beyond simple recall, the system incorporates cryptographic attestation, a coordination substrate with action DAGs and leases for multi-agent swarms, lifecycle management, and fine-grained governance to prevent context drift and ensure long-term consistency.