Home About Expertise Projects Research Blog Contact
Project Detail

Phoenix Workspace

In Development

AI-Augmented Knowledge Management

Unified system for knowledge management with living RAG, hybrid search, OCR pipeline, and Claude Code MCP integration for AI-assisted development.

Architecture Overview

Unified knowledge platform built around living RAG (retrieval-augmented generation) that continuously updates its knowledge base. Combines hybrid search using vector embeddings and SQLite FTS5 full-text search for maximum recall and precision. Includes an OCR pipeline for ingesting scanned documents and images into the knowledge graph. Integrates with Claude Code via the Model Context Protocol (MCP) for AI-assisted development and knowledge retrieval.

Key Features

Living RAG

Continuously updated retrieval-augmented generation

Hybrid Search

Vector embeddings + SQLite FTS5 full-text search

OCR Pipeline

Document and image ingestion via optical character recognition

MCP Integration

Claude Code Model Context Protocol for AI-assisted workflows

Memory System

Persistent knowledge graph with observation tracking

Tech Stack

Python FastAPI React TypeScript SQLite FTS5 Claude API MCP OCR

Screenshots

Screenshots coming soon