About Cortex Server
Cortex Server is an MCP server published by Yeseh in the AI category: mCP (Model Context Protocol) server for Cortex. This package enables AI agents to interact with the Cortex memory system through standardized tools and resources.
Six months or more without a commit doesn't mean the server is broken, but check the open issues before depending on it in production.
Install
claude mcp add cortex-server -- npx -y @yeseh/cortex-servernpx -y @yeseh/cortex-serverThis server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.
Ask AI
Ask AI about Cortex Server
Powered by Claude · Grounded in docs
Security checks
- ✓README presentThe repository ships documentation.
5 checks have not been run yet.
README
About
⚠️ Heads up: this project is experimental and under active development. No guarantees it will work as expected.
Cortex gives AI coding agents persistent, structured memory across sessions. Agents read and write named memories organized into stores and categories — so context survives between conversations, projects, and machines.
Packages
This is a monorepo containing the following packages:
| Package | Description |
|---|---|
@yeseh/cortex-core |
Core domain logic, types, and validation |
@yeseh/cortex-storage-fs |
Filesystem storage adapter |
@yeseh/cortex-cli |
Command-line interface |
@yeseh/cortex-server |
MCP server for AI agent integration |
Install
Cortex packages are published to npm under the @yeseh scope.
CLI (global)
npm install -g @yeseh/cortex-cli
# OR
bun install -g @yeseh/cortex-cli
MCP server
npm install -g @yeseh/cortex-server
# OR
bun install -g @yeseh/cortex-server
Agent Skill
Cortex comes with a bundled skill for your coding agent. You can install it using
npx skills add yeseh/cortex --skill memory
#OR
bunx skills add yeseh/cortex --skill memory
Quickstart (CLI)
Use the CLI to initialize a store and create your first memory in under a minute.
# 1) Initialize global config + default store
cortex init
# 2) (Optional) Initialize a project-local store in the current repo
cortex store init
# 3) Add a memory
cortex memory add notes/first-memory -c "Cortex is set up"
# 4) List memories
cortex memory list
# 5) Read a memory
cortex memory show notes/first-memory
For command reference and advanced usage, see CLI Guide.
Documentation
- CLI Guide
- MCP Server Guide
- Library Guide
- Configuration Reference
- Agent Instructions
- Memory Lifecycle
- Memory Organization
MCP Server
The MCP server exposes Cortex tools for agents over HTTP.
POST /mcp— MCP protocol endpointGET /health— health check
See MCP Server Guide for installation, startup, and client integration examples.
Development
# Install dependencies
bun install
# Run all tests
bun test packages
# Run tests for a specific package
bun test packages/core
bun test packages/cli
bun test packages/server
bun test packages/storage-fs
# Type check
bun run typecheck
# Lint
bun run lint
# Build all packages
bun run build
Project Structure
cortex/
├── assets/ # Logo and static assets
├── bin/ # Binary entry points
├── docs/ # User documentation
├── packages/
│ ├── core/ # @yeseh/cortex-core
│ │ └── src/
│ │ ├── memory/ # Memory types and operations
│ │ ├── category/ # Category and index management
│ │ ├── store/ # Store registry
│ │ ├── storage/ # Storage port interfaces
│ │ └── testing/ # Test helpers and mock adapters
│ ├── storage-fs/ # @yeseh/cortex-storage-fs
│ │ └── src/
│ │ └── ... # Filesystem storage adapter
│ ├── cli/ # @yeseh/cortex-cli
│ │ └── src/
│ │ ├── commands/ # CLI commands
│ │ └── run.ts # Entry point
│ └── server/ # @yeseh/cortex-server
│ └── src/
│ ├── memory/ # Memory MCP tools
│ ├── category/ # Category MCP tools
│ ├── store/ # Store MCP tools
│ └── index.ts # MCP server entry
├── scripts/ # Build and utility scripts
├── skills/ # Pre-built agent skills
├── openspec/ # Specifications and change proposals
├── package.json # Workspace root
└── tsconfig.base.json # Shared TypeScript config
License
MIT
README copy captured when this server was indexed; it may lag the repository. The original is authoritative.