Getting Started with Moltler

Choose Your Installation Path

🚀

Quick Start

Try Moltler in 5 minutes. Runs a local Elasticsearch cluster with everything pre-configured.

Best for: Demo, evaluation, development

Get started →

🏢

Existing Cluster

Install the plugin on your production Elasticsearch cluster. Download pre-built releases.

Best for: Production, enterprise

Install on existing cluster →

🤖

Connect to Cursor

Add Moltler skills to Cursor IDE. Your AI assistant gets access to all 180+ skills.

Best for: AI-assisted workflows

Connect Cursor →

Quick Start (5 Minutes)

One Command Setup

git clone --recurse-submodules https://github.com/bahaaldine/moltler.git
cd moltler && ./scripts/quick-start.sh

Builds the plugin, starts Elasticsearch, and you're ready to go.

Install Skills

cd hub && ./moltler-cli.sh install --all

Installs all 155+ skills into your cluster

Run Your First Skill

curl -u elastic-admin:elastic-password http://localhost:9200/_escript \
  -H "Content-Type: application/json" \
  -d '{"query": "RUN SKILL get_recent_errors()"}'

Install on Existing Cluster

1. Download the Plugin

Download the pre-built plugin zip that matches your Elasticsearch version:

📦 View Releases on GitHub →

Available versions:

ES 9.4.0ES 9.3.0ES 9.2.0ES 9.1.0ES 9.0.0ES 8.17.0ES 8.16.0ES 8.15.0

2. Install on Each Node

# Download the plugin (example for ES 9.4.0)
wget https://github.com/bahaaldine/moltler/releases/download/v1.0.0/elastic-script-9.4.0.zip

# Install the plugin
sudo /usr/share/elasticsearch/bin/elasticsearch-plugin install \
  file:///path/to/elastic-script-9.4.0.zip

# Restart Elasticsearch
sudo systemctl restart elasticsearch

Repeat on each node in your cluster

3. Configure CLI for Your Cluster

git clone https://github.com/bahaaldine/moltler.git
cd moltler/hub

# Configure for your cluster
export ES_URL="https://your-cluster:9200"
export ES_USER="elastic"
export ES_PASSWORD="your-password"

# Install skills
./moltler-cli.sh install --all

Architecture

┌─────────────────────────────────────────────────────────────────┐
│                        USER / AI AGENT                          │
└─────────────────────────────────────────────────────────────────┘
                              │
        ┌─────────────────────┼─────────────────────┐
        ▼                     ▼                     ▼
┌───────────────┐    ┌───────────────┐    ┌───────────────┐
│  MoltlerHub   │    │  Moltler CLI  │    │  Moltler MCP  │
│  (Web Portal) │    │  (Terminal)   │    │  (AI Bridge)  │
└───────────────┘    └───────────────┘    └───────────────┘
        │                     │                     │
        └─────────────────────┼─────────────────────┘
                              ▼
┌─────────────────────────────────────────────────────────────────┐
│               Elasticsearch + elastic-script plugin             │
│                        (Skills Runtime)                         │
└─────────────────────────────────────────────────────────────────┘

elastic-script plugin

Runs inside Elasticsearch. Executes skills on your data.

Required

Moltler CLI

Command-line tool to install and manage skills.

Required

MoltlerHub

Web portal to browse, search, and discover skills.

Optional

Moltler MCP

Bridge for AI agents (Claude, Cursor, etc.) to use skills.

Optional

Connect to Cursor IDE

One config, all skills. When you connect Cursor to Moltler, your AI assistant automatically gets access to all 180+ skills. No need to install skills individually.

Step 1: Create MCP Config File

Create .cursor/mcp.json in your project root (or global config):

{
  "mcpServers": {
    "moltler": {
      "command": "python",
      "args": ["/path/to/moltler/mcp-bridge/moltler_mcp_server.py"],
      "env": {
        "ES_URL": "http://localhost:9200",
        "ES_USER": "elastic-admin",
        "ES_PASSWORD": "elastic-password"
      }
    }
  }
}

Update the path and credentials for your environment.

Step 2: Install Dependencies

pip install httpx

Step 3: Restart Cursor

Restart Cursor IDE. The Moltler MCP server will appear in your MCP servers list. Your AI assistant now has access to all skills.

Alternative: Direct HTTP (if supported)

If your MCP client supports HTTP endpoints directly:

{
  "mcpServers": {
    "moltler": {
      "url": "http://localhost:9200/_escript/mcp",
      "headers": {
        "Authorization": "Basic ZWxhc3RpYy1hZG1pbjplbGFzdGljLXBhc3N3b3Jk"
      }
    }
  }
}

What Your AI Can Do

The Magic Moment

You:"I got paged for a production incident. Payment service seems down."

Agent:Uses get_recent_errors skill...

"I found 47 errors in payment-service. Most are 'Connection timeout to gateway'. Let me check transaction latency..."

Agent:Uses get_slow_transactions skill...

"Payment gateway calls averaging 12s (10x normal). Issue started at 9:45 AM. Want me to correlate with recent deployments?"

Your AI assistant can now investigate your Elasticsearch data like a senior engineer.

🔍

Observability

Query logs, analyze errors, find slow services

🛡️

Security

Detect threats, investigate incidents, manage alerts

⚙️

Cluster Ops

Health checks, index management, ILM policies

🎯

Kibana

Dashboards, alerts, cases, Fleet, ML jobs

🤖

AI/LLM

Summarize, classify, generate with inference API

🔧

Runbooks

Slack, PagerDuty, AWS, Kubernetes, Terraform

Works with: Cursor, Claude Desktop, Cline, Windsurf, and any MCP-compatible agent.