Quick Start Guide

Quick Start Guide

Get the MCP Context Provider running in 5 minutes with this step-by-step guide.

Overview

The MCP Context Provider gives Claude persistent memory across chat sessions. Instead of losing context when you restart Claude Desktop, your tool preferences, syntax rules, and best practices persist automatically.

Prerequisites

• Python 3.8+ (check with python --version)
• Claude Desktop installed and working
• Internet connection for downloading dependencies

Installation Methods

🚀 Option 1: Automated Installation (Recommended)

For Linux/macOS/Unix:

# One-line install
curl -sSL https://raw.githubusercontent.com/doobidoo/MCP-Context-Provider/main/install.sh | bash

For Windows:

# Download and run installer
Invoke-WebRequest -Uri "https://raw.githubusercontent.com/doobidoo/MCP-Context-Provider/main/install.bat" -OutFile "install.bat"
.\install.bat

What it does:

• ✅ Downloads the MCP Context Provider package
• ✅ Creates a Python virtual environment
• ✅ Installs all dependencies
• ✅ Configures Claude Desktop automatically
• ✅ Verifies the installation

Skip to Step 4 if using automated installation.

🔧 Option 2: Manual Installation

Step 1: Install Prerequisites

Install DXT CLI:

npm install -g @anthropic-ai/dxt

Verify installation:

dxt --version
python --version  # Should be 3.8+

Step 2: Download and Extract

# Download the DXT package
wget https://github.com/doobidoo/MCP-Context-Provider/raw/main/mcp-context-provider-1.2.1.dxt

# Extract to your preferred location
dxt unpack mcp-context-provider-1.2.1.dxt ~/mcp-context-provider
cd ~/mcp-context-provider

Step 3: Set Up Python Environment

# Create virtual environment
python -m venv venv

# Activate it
source venv/bin/activate  # Linux/macOS
# OR
venv\Scripts\activate     # Windows

# Install dependencies
pip install mcp>=1.9.4

Step 4: Configure Claude Desktop

Find your configuration file:

• Linux: ~/.config/claude/claude_desktop_config.json
• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

Add this configuration:

{
  "mcpServers": {
    "context-provider": {
      "command": "/home/user/mcp-context-provider/venv/bin/python",
      "args": ["/home/user/mcp-context-provider/server/context_provider_server.py"],
      "env": {
        "CONTEXT_CONFIG_DIR": "/home/user/mcp-context-provider/contexts",
        "AUTO_LOAD_CONTEXTS": "true"
      }
    }
  }
}

Important: Replace /home/user/mcp-context-provider with your actual installation path.

Step 5: Verify Installation

Run the verification script:

cd ~/mcp-context-provider
python verify_install.py

Expected output:

✓ Python version: 3.11.5
✓ MCP package installed
✓ Contexts directory found
✓ Found 5 JSON files
✓ Server loads successfully (5 contexts)
=== Installation verified ===

Step 6: Restart and Test

1. Restart Claude Desktop completely
2. Start a new chat in Claude Desktop
3. Test the installation by asking Claude:

Can you list available contexts using the MCP Context Provider?

Expected response: Claude should use the list_available_contexts tool and show:

• dokuwiki
• terraform
• azure
• git
• general_preferences

What You Get

Once installed, these tools are available in every Claude chat:

• list_available_contexts - See all loaded context categories
• get_context_rules - Get rules for specific tools (e.g., Terraform, Git)
• apply_auto_corrections - Auto-correct syntax (Markdown → DokuWiki)
• get_tool_preferences - Get your saved preferences for tools

Example Usage

Get Terraform naming rules:

Get the context rules for terraform

Auto-convert Markdown to DokuWiki:

Apply auto-corrections to this markdown for dokuwiki:
# My Header
This is `code` and a [link](http://example.com)

Check what contexts are available:

List all available contexts

Troubleshooting

Installation Issues

Problem: dxt: command not found

# Install Node.js first, then:
npm install -g @anthropic-ai/dxt

Problem: ModuleNotFoundError: No module named 'mcp'

# Activate venv and install:
source venv/bin/activate
pip install mcp>=1.9.4

Problem: Context Provider not showing in Claude

1. Check Claude Desktop config file syntax with: python -m json.tool claude_desktop_config.json
2. Verify all file paths in config are absolute and correct
3. Restart Claude Desktop completely

Quick Diagnostic

Run this to check your setup:

# Test server manually
cd ~/mcp-context-provider
source venv/bin/activate
python server/context_provider_server.py
# Should start without errors

Reset Installation

If something goes wrong:

# Remove installation
rm -rf ~/mcp-context-provider

# Remove from Claude config
# Edit your Claude Desktop config file:
# - Linux: ~/.config/claude/claude_desktop_config.json
# - macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
# - Windows: %APPDATA%\Claude\claude_desktop_config.json
# Remove the "context-provider" section

# Start over with fresh install

Advanced Configuration

Custom Context Location

Set a different context directory:

{
  "env": {
    "CONTEXT_CONFIG_DIR": "/path/to/my/custom/contexts"
  }
}

Debug Mode

Enable detailed logging:

{
  "env": {
    "DEBUG_MODE": "true"
  }
}

Multiple Environments

Use different contexts for different projects:

{
  "mcpServers": {
    "context-provider-work": {
      "command": "/path/to/venv/bin/python",
      "args": ["/path/to/server/context_provider_server.py"],
      "env": {
        "CONTEXT_CONFIG_DIR": "/work/contexts"
      }
    },
    "context-provider-personal": {
      "command": "/path/to/venv/bin/python", 
      "args": ["/path/to/server/context_provider_server.py"],
      "env": {
        "CONTEXT_CONFIG_DIR": "/personal/contexts"
      }
    }
  }
}

Next Steps

1. Explore the context files in the contexts/ directory
2. Customize existing contexts or create new ones
3. Read the Context Guide for advanced usage
4. Check Examples for real-world scenarios

Getting Help

• Installation issues: Check TROUBLESHOOTING.md
• Context customization: Read DEVELOPER_GUIDE.md
• Bug reports: Open an issue on GitHub

That's it! Your Claude Desktop now has persistent context that survives across all chat sessions. No more re-explaining your preferences every time you start a new conversation.