Skip to content

bhasinagam/ContextBridge

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

2 Commits
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

🧠 CodeContext

AI Coding Context Layer β€” Make AI assistants understand your team's codebase patterns

License: MIT TypeScript Next.js MCP Compatible

Features β€’ Quick Start β€’ Architecture β€’ API Reference β€’ Contributing


🎯 The Problem

AI coding assistants like Cursor and Copilot are powerful, but they don't understand your team's unique patterns:

  • Your error handling conventions
  • Your import style preferences
  • Your validation approach
  • Your authentication patterns

Every suggestion requires mental translation to match your codebase.

✨ The Solution

CodeContext automatically detects your codebase patterns and injects them as context into AI assistants via the Model Context Protocol (MCP). Now every AI suggestion follows your team's conventions out of the box.


πŸš€ Features

8 Pattern Types Detected

Category Pattern Types Detection Method
Code Style Error Handling, Imports, Naming, Structure Rules-based (instant)
Frameworks API Format, Validation, Database, Auth Hybrid (rules + LLM fallback)

Key Capabilities

  • πŸ” AST-Powered Analysis β€” Deep code parsing with Babel for accurate pattern detection
  • πŸ€– Hybrid Detection β€” Rules-based first, LLM (Groq) fallback for complex patterns
  • ⚑ Real-time Indexing β€” Background processing with Redis queue
  • πŸ”— MCP Protocol β€” Native integration with Cursor, Copilot, and more
  • πŸ” GitHub OAuth β€” Seamless onboarding and repository access
  • πŸ“Š Dashboard β€” View detected patterns, confidence scores, and usage analytics

πŸ—οΈ Architecture

flowchart TB
    subgraph Input["πŸ“₯ Input"]
        GH[GitHub Repository]
    end
    
    subgraph Core["🧠 CodeContext Engine"]
        IDX[Indexing Queue<br/>Redis/Upstash]
        AST[AST Parser<br/>Babel]
        DET[Pattern Detectors<br/>8 Types]
        LLM[LLM Fallback<br/>Groq]
        CTX[Context Generator]
    end
    
    subgraph Storage["πŸ’Ύ Storage"]
        DB[(PostgreSQL<br/>Supabase)]
    end
    
    subgraph Output["πŸ“€ Output"]
        MCP[MCP Server]
        DASH[Dashboard]
    end
    
    subgraph Clients["πŸ€– AI Clients"]
        CUR[Cursor]
        COP[Copilot]
        OTHER[Other MCP Clients]
    end
    
    GH --> IDX
    IDX --> AST
    AST --> DET
    DET --> LLM
    DET --> DB
    LLM --> DB
    DB --> CTX
    CTX --> MCP
    DB --> DASH
    MCP --> CUR
    MCP --> COP
    MCP --> OTHER
Loading

πŸ“¦ Tech Stack

Layer Technology
Frontend Next.js 14 (App Router), TypeScript, Tailwind CSS
Backend Next.js API Routes, NextAuth.js
Database PostgreSQL (Supabase)
Queue Redis (Upstash)
AST Parsing @babel/parser, @babel/traverse
LLM Groq (llama-3.3-70b-versatile)
MCP @modelcontextprotocol/sdk

πŸš€ Quick Start

Prerequisites

  • Node.js 18+
  • PostgreSQL database (we recommend Supabase)
  • Redis instance (we recommend Upstash)
  • GitHub OAuth App
  • Groq API key (free tier: 14,400 req/day)

1. Clone & Install

git clone https://github.com/bhasinagam/ContextBridge.git
cd ContextBridge
npm install

2. Configure Environment

cp .env.example .env.local

Edit .env.local with your credentials:

Variable Description Where to Get
DATABASE_URL PostgreSQL connection string Supabase β†’ Settings β†’ Database
UPSTASH_REDIS_REST_URL Redis REST URL Upstash β†’ Redis β†’ REST API
UPSTASH_REDIS_REST_TOKEN Redis REST token Same as above
NEXTAUTH_SECRET Random 32-byte secret Run: openssl rand -base64 32
GITHUB_CLIENT_ID OAuth App client ID GitHub β†’ OAuth Apps
GITHUB_CLIENT_SECRET OAuth App secret Same as above
GROQ_API_KEY Groq API key Groq Console

3. Set Up Database

Run the schema in your Supabase SQL editor:

-- Contents of src/lib/db/schema.sql

Or use the Supabase dashboard to import src/lib/db/schema.sql.

4. Run Development Server

npm run dev

Open http://localhost:3000 to access the dashboard.


πŸ”Œ MCP Integration

Using with Cursor

Add to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "codecontext": {
      "url": "http://localhost:3000/api/mcp/context",
      "headers": {
        "X-API-Key": "your-api-key"
      },
      "defaultParams": {
        "repo_id": "your-repo-id"
      }
    }
  }
}

API Usage

curl -X POST http://localhost:3000/api/mcp/context \
  -H "Content-Type: application/json" \
  -H "X-API-Key: your-api-key" \
  -d '{
    "query": "add an API endpoint",
    "repo_id": "your-repo-id"
  }'

πŸ“ Project Structure

src/
β”œβ”€β”€ app/                      # Next.js App Router
β”‚   β”œβ”€β”€ api/                  # API Routes
β”‚   β”‚   β”œβ”€β”€ auth/             # NextAuth endpoints
β”‚   β”‚   β”œβ”€β”€ github/           # GitHub API proxy
β”‚   β”‚   β”œβ”€β”€ mcp/              # MCP context & health
β”‚   β”‚   β”œβ”€β”€ patterns/         # Pattern queries
β”‚   β”‚   β”œβ”€β”€ repos/            # Repository CRUD
β”‚   β”‚   └── webhooks/         # GitHub webhooks
β”‚   β”œβ”€β”€ dashboard/            # Dashboard pages
β”‚   └── onboarding/           # Onboarding flow
β”œβ”€β”€ components/               # React components
β”‚   β”œβ”€β”€ ui/                   # shadcn/ui components
β”‚   └── dashboard/            # Dashboard-specific
└── lib/                      # Core libraries
    β”œβ”€β”€ db/                   # Database client & schema
    β”œβ”€β”€ github/               # GitHub API client
    β”œβ”€β”€ indexing/             # AST parser & queue
    β”œβ”€β”€ mcp/                  # MCP server & context generator
    β”œβ”€β”€ patterns/             # 8 pattern detectors
    └── utils/                # Types, helpers, Groq client

πŸ” Pattern Types

Rules-Based (No API Calls)

Pattern What It Detects
Error Handling try-catch blocks, wrapper functions (handleError, etc.)
Import Style Relative (./path), Absolute (@/, ~/), Barrel exports
Naming Convention camelCase, snake_case, PascalCase
File Structure App Router, Pages Router, Components directory

Hybrid (Rules + LLM Fallback)

Pattern What It Detects
API Format Next.js API Routes, response structure patterns
Validation Zod, Yup, Joi, Valibot, custom validation
Database Prisma, Drizzle, TypeORM, Mongoose, Supabase, Kysely
Authentication NextAuth, Clerk, Auth0, Supabase Auth, Firebase

πŸ”§ Troubleshooting

Database connection failed
  • Ensure your Supabase project is active
  • Check if the password contains special characters (URL-encode them)
  • Use port 5432 for direct connection, 6543 for pooled
GitHub OAuth redirect error
  • Verify callback URL is set to http://localhost:3000/api/auth/callback/github
  • Ensure NEXTAUTH_URL matches your app URL
Pattern detection returns empty
  • Check if repository indexing is complete (status: "completed")
  • Verify there are TypeScript/JavaScript files in the repo
  • Check Redis queue for pending jobs
MCP not connecting to Cursor
  • Restart Cursor after updating mcp.json
  • Check API key is valid and not expired
  • Verify the repo_id matches a indexed repository

πŸ—ΊοΈ Roadmap

  • Multi-language support β€” Python, Go, Rust
  • Custom pattern definitions β€” User-defined pattern rules
  • Team collaboration β€” Shared pattern configs
  • VS Code extension β€” Native VS Code integration
  • Pattern analytics β€” Usage trends and insights

🀝 Contributing

We welcome contributions! See CONTRIBUTING.md for guidelines.

  1. Fork the repository
  2. Create your feature branch (git checkout -b feature/amazing-feature)
  3. Commit your changes (git commit -m 'Add amazing feature')
  4. Push to the branch (git push origin feature/amazing-feature)
  5. Open a Pull Request

πŸ“„ License

This project is licensed under the MIT License - see the LICENSE file for details.


Built with ❀️ for the AI-assisted coding community

⭐ Star this repo if you find it useful!

About

AI Coding Context Layer - MCP server that makes AI assistants understand your team's codebase patterns

Topics

Resources

License

Code of conduct

Contributing

Security policy

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors