Content
# MCP AI Collab - Give AI Assistants Persistent Memory 🧠
**Transform stateless AI assistants into intelligent collaborators with persistent, project-aware memory.**
[](https://opensource.org/licenses/MIT)
[](https://www.python.org/downloads/)
[](https://github.com/anthropics/mcp)
## 🎯 The Problem We Solve
When you use AI assistants like Gemini, Grok, or ChatGPT through Claude Code, they forget everything between calls:
```
You: "Hey Gemini, help me debug this auth function"
Gemini: "I found the issue on line 42..."
// 5 minutes later...
You: "Gemini, what was that issue we found?"
Gemini: "I don't have any previous context..." 😔
```
**This MCP server fixes that!** Each AI now remembers your conversations per project:
```
You: "Gemini, what was that issue we found?"
Gemini: "We found a race condition in the auth function on line 42.
We discussed using a mutex lock to fix it." 🎯
```
## 🌟 Key Features
- **🧠 Persistent Memory** - Each AI maintains conversation history per project
- **📁 Project Isolation** - Separate contexts for different projects automatically
- **🚀 Three Versions** - Choose based on your needs (see below)
- **🔒 100% Secure** - Your API keys never leave your machine
- **⚡ Fast Performance** - Redis caching with PostgreSQL persistence
- **🔧 Easy Setup** - One-click installation with interactive menu
## 📋 Choose Your Version
We offer three versions to match your needs:
| Version | Storage | Setup Time | Best For |
|---------|---------|------------|----------|
| **Clean** (Recommended) | Local JSON files | 1 minute | Individual developers, quick start |
| **Full** | Redis + PostgreSQL | 5 minutes | Teams, production, high performance |
| **Standalone** | Local JSON files | 1 minute | Learning MCP, minimal dependencies |
See [detailed comparison](docs/VERSION_COMPARISON.md) →
## 🚀 Quick Start
```bash
# Clone the repository
git clone https://github.com/RaiAnsar/claude_code-coding-mcp.git
cd claude_code-coding-mcp
# Run interactive setup (recommended)
./one_click_setup.sh
```
The setup wizard will:
1. Check your system requirements
2. Let you choose a version
3. Guide you through API key configuration
4. Install everything automatically
## 🔑 Configuration
### Getting API Keys (Required)
You'll need at least one API key:
- **Gemini**: [Google AI Studio](https://makersuite.google.com/app/apikey) (Free tier available)
- **Grok**: [X.AI Console](https://console.x.ai/) (Requires X Premium+)
- **ChatGPT**: [OpenAI Platform](https://platform.openai.com/api-keys) (Pay as you go)
### Setting API Keys
During setup, you'll be prompted to enter your keys. They're stored in a local `.env` file:
```env
GEMINI_API_KEY=your-gemini-key-here
GROK_API_KEY=your-grok-key-here
OPENAI_API_KEY=your-openai-key-here
```
### Customizing AI Models
You can customize which models to use for each AI service in your `.env` file:
```bash
# Default models (latest versions)
GEMINI_MODEL=gemini-2.5-pro-preview-06-05
GROK_MODEL=grok-3
OPENAI_MODEL=gpt-4o
# You can change to other available models:
# GEMINI_MODEL=gemini-2.0-flash-001
# OPENAI_MODEL=gpt-4o
# GROK_MODEL=grok-2
```
Simply edit your `.env` file to use different models based on your needs (cost, performance, capabilities).
## 📖 How to Use
After installation and restarting Claude Code:
### Basic Commands
```bash
# Check if everything is working
Use db_status
# Ask AIs questions (they'll remember context)
Use ask_gemini to explain this authentication flow
Use ask_grok to help optimize this algorithm
Use ask_openai to review our API design
# Check conversation history
Use show_context with ai "gemini"
# Clear memory for fresh start
Use clear_context with ai "all"
```
### Real Example Session
```
You: Use ask_gemini to analyze the performance bottleneck in our API
Gemini: I can see the main bottleneck is in the database query on line 234...
You: Use ask_gemini to suggest optimization strategies
Gemini: Based on our previous analysis of the bottleneck on line 234,
here are three optimization strategies...
[Gemini remembers the context!]
```
## 🏗️ Architecture
```mermaid
graph TD
A[Claude Code] -->|MCP Protocol| B[MCP AI Collab Server]
B --> C{Router}
C --> D[Gemini API]
C --> E[Grok API]
C --> F[OpenAI API]
B --> G[Context Manager]
G --> H[Redis Cache]
G --> I[PostgreSQL]
G --> J[Local JSON]
style A fill:#f9f,stroke:#333,stroke-width:4px
style B fill:#bbf,stroke:#333,stroke-width:4px
style G fill:#bfb,stroke:#333,stroke-width:4px
```
## 🔒 Security & Privacy
**Your API keys are 100% safe:**
- ✅ **Local Storage Only** - Keys are stored in `.env` file on your machine
- ✅ **Never Transmitted** - Except to official AI APIs (Google, X.AI, OpenAI)
- ✅ **Never Logged** - No keys in logs, console output, or error messages
- ✅ **Gitignored** - `.env` files are excluded from version control
- ✅ **Open Source** - Review our code anytime
[Read our Security Policy](SECURITY.md) →
## 📚 Documentation
- [Quick Start Guide](docs/QUICKSTART.md) - Get running in 5 minutes
- [Architecture Overview](docs/ARCHITECTURE.md) - Technical deep dive
- [Version Comparison](docs/VERSION_COMPARISON.md) - Choose the right version
- [Troubleshooting](docs/TROUBLESHOOTING.md) - Common issues and solutions
- [Examples](docs/EXAMPLES.md) - Real-world usage patterns
## ❓ FAQ
**Q: Is this trying to steal my API keys?**
A: No! Your keys stay on your machine. Check our source code - we're fully open source.
**Q: Which version should I use?**
A: Start with the Clean version. You can always upgrade to Full later.
**Q: Can I use just one AI?**
A: Yes! You only need API keys for the AIs you want to use.
**Q: Does this work with Claude's /clear command?**
A: Yes! When you clear Claude's context, it clears the AI contexts too.
**Q: How is this different from using AIs directly?**
A: This gives them memory within Claude Code, making them true collaborators.
## 🤝 Contributing
We welcome contributions! See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines.
## 📜 License
MIT License - see [LICENSE](LICENSE) for details.
## 🙏 Acknowledgments
Built on top of the [MCP protocol](https://github.com/anthropics/mcp) by Anthropic.
---
**Ready to give your AI assistants persistent memory?** [Get started now!](#-quick-start) 🚀
