Cortex
Local-first, end-to-end-encrypted memory engine for AI agents that provides persistent, privacy-preserving memory across sessions, using Rust and MCP.
Cortex
What is Cortex?
How to use Cortex?
- 1Install the cortex-mcp-server via Homebrew (brew install gambletan/tap/cortex-mcp-server) or build from source with Cargo.
- 2Run the server with a database path: cortex-mcp-server ~/.cortex/memory.db.
- 3Register as an MCP server for Claude Code or Claude Desktop.
- 4Use the 30 MCP tools to store, search, and manage memories.
- 5Optionally enable cloud sync for cross-device memory.
Cortex Key Features
- Private by default (local-first, zero telemetry)
- 4-tier memory model (Working, Episodic, Semantic, Procedural)
- Sub-millisecond ingest (156µs) and search (568µs)
- 30 MCP tools for Claude and MCP clients
- Cross-device sync via iCloud/Drive/Dropbox with AES-256-GCM encryption
- Bayesian belief system with self-correction
- People graph with cross-channel identity resolution
- Open-source (MIT license)
Cortex Use Cases
- Give AI persistent memory across sessions
- Personal AI assistant that remembers user preferences and facts
- Cross-device memory for Claude Desktop and Claude Code
- Automatic extraction of facts, relationships, and beliefs from conversations
- Multi-project isolation for developers working with multiple codebases
Cortex Pricing & Free Credits
Cortex currently operates on a Free model.
Cortex Pros & Cons
Pros
- 100% local, no data sent to third parties
- Fast performance (sub-millisecond operations)
- Rich feature set (4 memory tiers, beliefs, people graph)
- Free and open-source with no usage limits
- Easy integration with MCP-compatible clients
Cons
- Requires Rust setup if not using Homebrew
- One-time model download for semantic search (~30MB)
- Limited to MCP client ecosystem
- No built-in mobile app (browser demo available)
What is Cortex best for?
- Developers building AI agents and assistants
- Users of Claude Desktop seeking persistent memory
- Privacy-conscious individuals who want local AI memory
- Teams requiring cross-device memory sync without cloud dependency
Cortex FAQ
Top free alternatives to Cortex
A multi-agent AI runtime for Python developers with OS-inspired primitives like job scheduling, DAG orchestration, memory, and tool execution.
Memento is an AI tool that surfaces your important history from email, organizing it around people, projects, and topics that matter to you.
Viktor is an AI employee that connects to over 3,200 tools and performs tasks like generating reports, dashboards, code, and campaigns directly in Slack or Microsoft Teams.
Happycapy is an agent-native computer platform that provides a cloud-based sandbox environment for running AI agents to automate tasks via a visual desktop interface.
Command Center is an agentic coding environment that helps teams ship AI-generated code to production 2x faster.
Zero.xyz gives AI agents instant access to over 4,000 tools, APIs, and services without accounts or API keys.
WorkBeaver is an agentic automation platform that automates repetitive administrative tasks directly within your existing tools without any coding.
Best alternatives AI Tools to Cortex
A multi-agent AI runtime for Python developers with OS-inspired primitives like job scheduling, DAG orchestration, memory, and tool execution.
Memento is an AI tool that surfaces your important history from email, organizing it around people, projects, and topics that matter to you.
ctx is a context-aware recommendation engine that dynamically selects the best skills, agents, and MCPs for your current coding task using a massive knowledge graph.
Local-first AI intelligence engine for macOS that builds private memory from screen OCR, email, and calendar, then provides proactive assistance and autonomous task execution.
Reality Kernel provides causal verification for AI agents by simulating actions across shadow worlds, preventing attacks undetectable by LLM judges.
A Go-native framework for building, orchestrating, and observing AI agents with native OpenTelemetry observability and a self-hosted dashboard.
AIsa provides a single API key for AI agents to access 1000+ LLMs, APIs, and skills with autonomous micropayments.