Cortex-Mem: A Production-Ready Memory Framework for Autonomous AI Systems

Introduction

Cortex-Mem is a production-ready, AI-native memory framework built in Rust, designed to provide autonomous systems with intelligent long-term memory. It moves beyond simple chat history, offering a sophisticated, hierarchical three-tier memory architecture, from extraction and search to automated optimization. This framework empowers AI agents, such as those in Embodied-AI and OpenClaw, with the ability to remember, learn, and personalize interactions across sessions.

At its core, Cortex-Mem utilizes a hybrid storage architecture, combining virtual-filesystem durability with high-performance vector-based semantic search. It comes with out-of-the-box support for SKILL, CLI, API, MCP, and an insights dashboard, making it a comprehensive solution for building truly intelligent, context-aware applications.

Why Use Cortex-Mem?

Cortex-Mem transforms stateless AI into an intelligent, context-aware partner. Before Cortex-Mem, AI often forgets user details, lacks personalization, and repeats questions, leading to robotic and impersonal interactions. With Cortex-Mem, AI agents remember user preferences and history, provide deeply personalized interactions, learn and adapt over time, and maintain context across multiple conversations. This allows them to build rapport and feel like a true assistant.

It is ideal for developers building LLM-powered chatbots and agents, teams creating personalized AI assistants, and open-source projects requiring a robust memory backbone. Cortex-Mem has also demonstrated superior performance and token efficiency in benchmarks, achieving significantly higher scores with fewer tokens compared to other systems.

Installation

To get started with Cortex-Mem, you will need:

• Rust (version 1.86 or later)
• Qdrant vector database (version 1.7+)
• An OpenAI-compatible LLM API endpoint for memory extraction and analysis
• An OpenAI-compatible Embedding API endpoint for vector search

The simplest way to install the core components is via cargo:

# Install the CLI for command-line management
cargo install --path cortex-mem-cli

# Install the REST API Service for application integration
cargo install --path cortex-mem-service

# Install the MCP server for AI assistant integrations
cargo install --path cortex-mem-mcp

Cortex-Mem applications are configured via a config.toml file, which specifies details for Qdrant, LLM, embedding services, and the data directory.

Examples

The cortex-mem-cli provides a powerful interface for direct interaction with the memory system. All commands require a config.toml file and can use the --tenant flag for multi-tenant isolation.

Add a Memory:

Adds a new message to a session thread.

cortex-mem --config config.toml --tenant acme add --thread thread-123 --role user "The user is interested in Rust programming."

Search for Memories:

Performs a semantic vector search across the memory store.

cortex-mem --config config.toml --tenant acme search "what are the user's hobbies?" --thread thread-123 --limit 10

List Memories:

Retrieves a list of memories from a specific URI path.

cortex-mem --config config.toml --tenant acme list --uri "cortex://session" --include-abstracts

Get a Specific Memory:

Retrieves a specific memory by its URI.

cortex-mem --config config.toml --tenant acme get "cortex://session/thread-123/memory-456.md"

Session Management:

Create and close conversation sessions. Closing a session triggers memory extraction and indexing.

# Create a new session
cortex-mem --config config.toml --tenant acme session create thread-456 --title "My Session"

# Close a session
cortex-mem --config config.toml --tenant acme session close thread-456

Links

• GitHub Repository
• Documentation
• MIT License