Concept Tracker
Automatically extracts technical concepts from AI coding conversations, organizes them into a searchable knowledge base with hierarchy and categories, and links them to specific locations in your codebase.
README
Concept Tracker
An MCP (Model Context Protocol) server that automatically extracts technical concepts from your AI coding conversations, organizes them into a searchable knowledge base, and links them to your actual codebase.
Overview
When working with AI coding assistants, you discuss countless technical concepts — libraries, design patterns, language features, architectural decisions. These valuable learning moments get buried in chat history and forgotten.
Concept Tracker captures this knowledge automatically. It hooks into your conversations in real-time, extracts technical concepts, and builds a per-project knowledge base that shows you:
- What concepts you've learned and discussed
- Why they matter (with explanations)
- Where they appear in your code
Features
Multi-IDE Support
- Claude Code: Native hook integration for automatic extraction
- Cursor: Full hook support with stop event handling
- Continue.dev: Webhook-based integration
- Universal VS Code extension works across all AI tools
Real-Time Concept Extraction
- Hooks into AI coding conversations as they happen
- LLM-powered extraction identifies technical concepts automatically
- Captures original chat context for future reference
Smart Organization
- Hierarchy: Concepts organized in parent-child relationships (e.g., "useState" under "React Hooks")
- Categories: Language features, libraries/frameworks, design patterns, architectural decisions
- Deduplication: Exact name matching prevents duplicate entries
Codebase Linking
- Real-time scanning on file save
- Finds where each concept appears in your project
- Direct links to specific file locations
Dual Dashboard
- IDE Panel: Quick access without leaving your editor
- Web App: Full-featured dashboard for deeper exploration
Knowledge Management
- Edit concept names and explanations
- Merge similar concepts
- Manual concept addition
- Export to JSON or Markdown
- Notifications when new concepts are extracted
Concept Structure
Each concept contains:
{
"id": "uuid",
"name": "useState",
"category": "library",
"parent": "React Hooks",
"explanation": "A React Hook that lets you add state to functional components...",
"chatSnippets": [
{
"timestamp": "2025-01-15T10:30:00Z",
"content": "useState returns a pair: the current state value and a function to update it..."
}
],
"codeLocations": [
"src/components/Counter.tsx:12",
"src/hooks/useAuth.ts:8"
],
"firstSeen": "2025-01-15T10:30:00Z",
"lastSeen": "2025-01-20T14:22:00Z"
}
Architecture
┌─────────────────────────────────────────────────────────────────┐
│ Claude Code │
│ │ │
│ (hooks API) │
│ ▼ │
│ ┌─────────────────────────────────────────────────────────┐ │
│ │ Concept Tracker MCP │ │
│ │ ┌───────────────┐ ┌───────────────┐ ┌─────────────┐ │ │
│ │ │ Extractor │ │ Hierarchy │ │ Scanner │ │ │
│ │ │ (LLM) │ │ Manager │ │ (Codebase) │ │ │
│ │ └───────────────┘ └───────────────┘ └─────────────┘ │ │
│ │ │ │ │
│ │ ┌──────▼──────┐ │ │
│ │ │ Storage │ │ │
│ │ │ (Local) │ │ │
│ │ └─────────────┘ │ │
│ └─────────────────────────────────────────────────────────┘ │
│ │ │
│ ┌─────────────┴─────────────┐ │
│ ▼ ▼ │
│ ┌─────────────────┐ ┌─────────────────┐ │
│ │ IDE Panel │ │ Web App │ │
│ │ (VS Code) │ │ (localhost) │ │
│ └─────────────────┘ └─────────────────┘ │
└─────────────────────────────────────────────────────────────────┘
Concept Categories
| Category | Examples |
|---|---|
| Language Features | async/await, generics, decorators, pattern matching |
| Libraries & Frameworks | React hooks, Express middleware, Prisma models |
| Design Patterns | Dependency injection, observer pattern, factory pattern |
| Architecture | Microservices, event sourcing, CQRS, hexagonal architecture |
Roadmap
Phase 1: MVP
- [x] Project setup
- [x] Basic concept extraction (DeepSeek API)
- [x] Local JSON storage
- [x] Simple web dashboard
- [x] Claude Code hook integration (auto-extract on conversation)
Phase 2: Enhanced Features
- [x] Hierarchy management UI
- [x] Real-time codebase scanning
- [x] VS Code panel integration
- [x] Concept merge/edit functionality
- [x] Export capabilities
Phase 3: Multi-IDE Support (Current)
- [x] Cursor integration
- [x] Continue.dev integration
- [x] Universal VS Code extension (works with any AI tool)
- [x] Unified configuration system
- [x] IDE adapter abstraction layer
Tech Stack
- MCP Server: TypeScript
- Storage: Local JSON/SQLite
- Web Dashboard: React + Vite
- IDE Panel: VS Code Webview API
- Code Scanning: Tree-sitter / ripgrep
Getting Started
Prerequisites
- Node.js 18+
- npm 9+
- DeepSeek API key (for concept extraction)
Installation
# Clone and enter the project
cd concept-tracker
# Install all dependencies
npm install
# Create your .env file
cp .env.example .env
# Edit .env and add your DEEPSEEK_API_KEY
# Build the MCP server
npm run build
# Run the universal installer (detects and configures all IDEs)
./scripts/install.sh
# Start the servers
npm run dev:api # API server (port 3001)
npm run dev # Dashboard (port 3000)
IDE-Specific Installation
If you prefer to install hooks for specific IDEs:
# Claude Code only
./scripts/install-claude-hook.sh
# Cursor only
./scripts/install-cursor-hook.sh
# Continue.dev only
./scripts/install-continue-hook.sh
Running the Dashboard
# Development mode with hot reload
npm run dev
# The dashboard will open at http://localhost:3000
Building for Production
# Build both MCP server and dashboard
npm run build
Configuration
Remote MCP Setup (Hosted Service)
If you're using a hosted version of Concept Tracker, add this to your Cursor MCP config (~/.cursor/mcp.json):
{
"mcpServers": {
"concept-tracker": {
"url": "https://your-deployed-url.railway.app/sse?token=YOUR_UNIQUE_TOKEN"
}
}
}
Replace YOUR_UNIQUE_TOKEN with a unique identifier (8-64 alphanumeric characters or dashes). This token isolates your concepts from other users.
Claude Code MCP Setup
Add to your Claude Code MCP configuration (~/.claude.json or project .claude/settings.json):
{
"mcpServers": {
"concept-tracker": {
"command": "node",
"args": ["/path/to/concept-tracker/mcp-server/dist/index.js"],
"env": {
"DEEPSEEK_API_KEY": "your-api-key-here"
}
}
}
}
Environment Variables
| Variable | Description | Required |
|---|---|---|
DEEPSEEK_API_KEY |
Your DeepSeek API key for concept extraction | Yes |
STORAGE_PATH |
Custom storage path (default: ~/.concept-tracker) |
No |
MCP Tools
The Concept Tracker MCP server provides these tools:
| Tool | Description |
|---|---|
extract_concepts |
Extract technical concepts from conversation text |
list_concepts |
List all concepts with optional category/search filters |
get_concept |
Get detailed info about a specific concept |
add_concept |
Manually add a new concept |
update_concept |
Update a concept's name or explanation |
delete_concept |
Remove a concept from the knowledge base |
Deploying to Railway (Self-Hosting)
To host your own public Concept Tracker MCP:
1. Prerequisites
- A Railway account
- This repository pushed to GitHub
2. Deploy
# Install Railway CLI
npm install -g @railway/cli
# Login to Railway
railway login
# Initialize project in this directory
railway init
# Link to your project
railway link
# Set your DeepSeek API key
railway variables set DEEPSEEK_API_KEY=your-api-key-here
# Deploy
railway up
Or use the Railway dashboard:
- Create a new project
- Connect your GitHub repo
- Add environment variable:
DEEPSEEK_API_KEY - Railway will auto-detect and deploy
3. Share with Users
Once deployed, share the URL with users. They'll configure Cursor like this:
{
"mcpServers": {
"concept-tracker": {
"url": "https://YOUR-APP.railway.app/sse?token=their-unique-token"
}
}
}
Each user should create their own unique token for isolated concept storage.
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
E2B
Using MCP to run code via e2b.