Skip to main content
Glama
billebel

Catalyst MCP Server

by billebel

Catalyst MCP Server

MCP (Model Context Protocol) server implementation that loads and serves Knowledge Packs.

Docker MCP Protocol Pack Builder License: MIT GitHub Stars

Features

  • MCP server implementation with FastAPI

  • LibreChat integration for web interface

  • Docker deployment support

  • Knowledge Pack loading from YAML configurations

  • Authentication and rate limiting

  • Support for multiple AI models (Claude, GPT, Gemini)

Quick Start

1. Clone and Configure

# Clone the repository
git clone https://github.com/billebel/catalyst_mcp.git
cd catalyst_mcp

# Copy environment template
cp .env.example .env

# Edit .env with your API keys
nano .env

2. Set Your API Keys

Edit .env file:

# Add your API keys
ANTHROPIC_API_KEY=your-claude-api-key
OPENAI_API_KEY=your-openai-api-key        # Optional
GOOGLE_API_KEY=your-gemini-api-key        # Optional

# JWT secrets (change for production!)
JWT_SECRET=your-secure-jwt-secret
JWT_REFRESH_SECRET=your-secure-refresh-secret

3. Start with Docker

# Start the complete stack
docker-compose up -d

# View logs
docker-compose logs -f

4. Access Your AI Assistant

Architecture

graph TD
    A[AI Assistant<br/>Claude Desktop] --> B[MCP Protocol]
    C[Web Chat<br/>LibreChat] --> B
    B --> D[Catalyst MCP Server]
    D --> E[Knowledge Packs]
    E --> F[Your Business Systems]
    F --> G[Databases]
    F --> H[REST APIs] 
    F --> I[Cloud Services]

Knowledge Packs

Catalyst includes example Knowledge Packs for common business systems:

Pack

Description

Use Cases

PostgreSQL Analytics

Database queries and reporting

Business intelligence, data analysis

GitHub DevOps

Repository management and CI/CD

Code management, deployment tracking

GitLab DevOps

GitLab API integration

Project management, pipeline monitoring

Linux Server Admin

Server management and monitoring

System administration, log analysis

RabbitMQ Messaging

Message queue management

Queue monitoring, message handling

S3 Storage

AWS S3 file operations

File management, backup operations

Creating Custom Packs

Create Knowledge Packs using the Catalyst Builder:

# Install the pack builder
pip install catalyst-builder

# Create a new CRM integration pack
catalyst-packs create crm-integration \
  --type rest \
  --description "Connect to our CRM system"

# This generates a complete pack structure:
# crm-integration/
# ├── pack.yaml           # Main configuration
# ├── tools/              # Tool definitions
# ├── prompts/            # AI prompts
# └── README.md           # Documentation

The generated pack.yaml:

metadata:
  name: crm-integration
  description: "Connect to our CRM system"
  domain: sales

connection:
  type: rest
  base_url: "https://api.yourcrm.com/v1"
  auth:
    method: bearer
    token: "${CRM_API_TOKEN}"

tools:
  - name: search_customers
    type: search
    description: "Find customers by name or email"
    endpoint: "/customers/search"

Pack Builder Resources:

Deployment Options

# Production deployment
docker-compose up -d

# Development with hot reload
docker-compose -f docker-compose.yml -f docker-compose.override.yml up -d

Local Development

# Install Python dependencies
pip install -r requirements.txt

# Start MCP server
python -m catalyst_mcp.server

# Start chat interface (separate terminal)
# See docs/chat-customization.md for LibreChat setup

Configuration

Environment Variables

Variable

Description

Required

Default

MCP_PORT

MCP server port

No

8443

MCP_HOST

Server bind address

No

0.0.0.0

LOG_LEVEL

Logging level

No

INFO

ANTHROPIC_API_KEY

Claude API key

Yes*

-

OPENAI_API_KEY

OpenAI API key

No

-

GOOGLE_API_KEY

Gemini API key

No

-

JWT_SECRET

Chat authentication

Yes

-

ALLOW_REGISTRATION

Allow new users

No

false

*At least one AI provider API key is required.

Chat Interface Customization

Catalyst uses LibreChat for the web interface. Customize:

  • Branding: Edit librechat.yaml for colors, logos

  • Authentication: Configure OAuth providers in .env

  • Models: Enable/disable AI models per user

  • Plugins: Add custom plugins and tools

See: Chat Customization Guide

AI Assistant Integration

Claude Desktop

Add to your Claude Desktop configuration:

{
  "mcpServers": {
    "catalyst": {
      "command": "mcp-client",
      "args": ["--url", "http://localhost:8443"]
    }
  }
}

ChatGPT/OpenAI

Use the MCP-compatible plugin or direct API integration.

Custom AI Applications

Connect any MCP-compatible AI application:

import mcp_client

# Connect to Catalyst MCP server
client = mcp_client.MCPClient("http://localhost:8443")

# Use business tools
result = client.call_tool("search_customers", {"query": "ACME Corp"})

Security Features

Authentication & Authorization

  • JWT-based session management

  • Role-based access control

  • OAuth provider integration (GitHub, Google, etc.)

API Security

  • Rate limiting and request throttling

  • Input validation and sanitization

  • Audit logging for compliance

Deployment Security

  • HTTPS/TLS encryption

  • Environment variable secrets

  • Container isolation

Examples & Use Cases

Business Intelligence

Use the Catalyst Builder to create database analytics packs:

catalyst-packs create bi-dashboard --type database --description "Executive dashboard"

DevOps Automation

Create deployment and monitoring packs:

catalyst-packs create devops-tools --type rest --description "CI/CD automation"

Customer Support

Build support system integrations:

catalyst-packs create support-tools --type rest --description "Help desk integration"

Community & Support

License

MIT License


Quick Commands

# Start everything
docker-compose up -d

# View logs
docker-compose logs -f catalyst-mcp

# Stop services
docker-compose down

# Create custom packs
pip install catalyst-builder
catalyst-packs create my-integration --type rest

Getting Started

  1. Clone repository: git clone https://github.com/billebel/catalyst_mcp.git

  2. Install pack builder: pip install catalyst-builder

  3. Create packs as needed

  4. Deploy with Docker: docker-compose up -d

F
license - not found
-
quality - not tested
C
maintenance

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/billebel/catalyst_mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server