corticai

CorticAI Documentation

Welcome to the CorticAI documentation. CorticAI is a comprehensive context engine that provides persistent, intelligent memory for development projects through a sophisticated multi-layer architecture.

What is CorticAI?

CorticAI is a TypeScript-based context management system built on the Mastra.ai framework. It acts as a “digital cortex” for your projects, maintaining context across sessions, preventing knowledge loss, and enabling AI agents and developers to understand code and documents from multiple perspectives.

Documentation Structure

Core Documentation

• Architecture Overview - System design and component relationships
• Getting Started - Installation and quick start guide
• API Reference - Complete API documentation

Component Guides

• Storage System - Storage adapters and persistence
• Query System - Query builder and execution
• Context Management - Context storage and retrieval
• Adapters - Domain and storage adapter patterns
• Indexes - Attribute-based indexing system

Integration Guides

• Mastra Integration - Using CorticAI with Mastra agents
• MCP Server - Model Context Protocol integration

Development

• Testing Guide - Running and writing tests
• Benchmarking - Performance measurement
• Contributing - Development guidelines

Quick Links

• Source Code: GitHub Repository
• API Docs: TypeDoc Reference
• Issues: Issue Tracker

Key Features

• Intelligent Memory: Automatic deduplication, validation, and enrichment of context
• Multi-Storage Support: Memory, JSON, and DuckDB storage adapters
• Natural Language Queries: Query stored context using natural language
• Type-Safe Query Builder: Fluent API with full TypeScript support
• High Performance: Optimized for large datasets with sub-100ms query times
• Extensible Architecture: Plugin-based design for custom adapters

Project Status

CorticAI is production-ready with:

• ✅ 759/759 tests passing (100% pass rate)
• ✅ Core components implemented and tested
• ✅ Mastra framework 0.16.3 integration
• ✅ Comprehensive API documentation
• ✅ Performance benchmarking suite

License

MIT License - See LICENSE for details.