> ## Documentation Index
> Fetch the complete documentation index at: https://docs.agent-corex.com/llms.txt
> Use this file to discover all available pages before exploring further.

# What is Agent-CoreX?

> Dynamic MCP-based AI agent framework that enables developers to build intelligent agents with access to 100+ MCP servers.

## Introduction to Agent-CoreX

Agent-CoreX is a powerful framework for building **AI agents that dynamically discover and execute tools** using the Model Context Protocol (MCP). It's designed for developers building intelligent automation workflows that need flexible, on-demand access to MCP servers in AI applications.

### Why Agent-CoreX?

Traditional AI agents are limited by a fixed set of pre-configured tools. Agent-CoreX changes this by introducing **dynamic tool retrieval** — your agent can ask for the right tools when it needs them, not before.

**Key Benefits:**

* **Dynamic Tool Discovery** - Find the perfect tool using natural language
* **MCP Servers for Developers** - Seamless integration with 100+ MCP servers in the ecosystem
* **Tool Ranking & Optimization** - Intelligent selection based on your use case
* **Zero Configuration** - Works out of the box with major MCP servers
* **Production-Ready** - Built for enterprise workflows and DevOps automation

### What Problems Does Agent-CoreX Solve?

**Before Agent-CoreX:**

```
Agent needs GitHub + Jira + Slack
↓
Hardcode 3 tools at startup
↓
Agent can only use these 3 tools
↓
Adding a new tool? Rebuild everything.
```

**With Agent-CoreX:**

```
Agent needs tools for current task
↓
Query /retrieve_tools with natural language
↓
Get ranked, relevant tools instantly
↓
Execute the right tool for the job
↓
Dynamically adapt to any workflow
```

### Core Features

<CardGroup cols={2}>
  <Card title="Tool Retrieval API" icon="search" href="/tool-retrieval/how-it-works">
    Query natural language, get ranked tools instantly. No manual selection.
  </Card>

  <Card title="MCP Server Integration" icon="plug" href="/mcp/setup-mcp-servers">
    Connect MCP servers for developers with built-in support for 100+ servers.
  </Card>

  <Card title="Smart Ranking" icon="star" href="/tool-retrieval/ranking-logic">
    AI-powered ranking ensures your agent always gets the most relevant tools.
  </Card>

  <Card title="Token Optimization" icon="zap" href="/tool-retrieval/optimization">
    Minimize token usage with intelligent caching and context optimization.
  </Card>
</CardGroup>

### How It Works (High Level)

```
┌─────────────────────────────────────────────────────────┐
│                    Your AI Agent                        │
└─────────────────┬───────────────────────────────────────┘
                  │ "I need to automate GitHub workflows"
                  ↓
┌─────────────────────────────────────────────────────────┐
│            Tool Retrieval API                           │
│        /retrieve_tools (Natural Language Query)         │
└─────────────────┬───────────────────────────────────────┘
                  │ Returns: ["create-pr", "deploy", "notify"]
                  ↓
┌─────────────────────────────────────────────────────────┐
│              MCP Manager                                │
│    Routes requests to appropriate MCP servers           │
└─────────────────┬───────────────────────────────────────┘
                  │
        ┌─────────┴──────────┬──────────┐
        ↓                    ↓          ↓
    ┌────────┐          ┌────────┐  ┌────────┐
    │ GitHub │          │ Slack  │  │ Jira   │
    │ MCP    │          │ MCP    │  │ MCP    │
    └────────┘          └────────┘  └────────┘
```

### Perfect For

* **DevOps Teams** - Automate infrastructure workflows with Terraform, Docker, and cloud CLIs
* **GitHub Automation** - MCP servers for developers enabling PR reviews, CI/CD, and deployments
* **Enterprise Workflows** - Connect Jira, Slack, and custom tools seamlessly
* **Multi-Tool Orchestration** - Coordinate across MCP servers in VS Code, Claude Code, and Copilot

### MCP Servers in AI Applications

Agent-CoreX is optimized for working with **MCP servers in AI** environments:

* **MCP Servers for Claude Code** - Native integration with Claude Code IDE extension
* **MCP Servers in VS Code** - Full support for VS Code MCP marketplace
* **MCP Servers for GitHub Copilot** - Seamless Copilot integration
* **MCP Servers in Copilot** - Enterprise Copilot workflows

### Next Steps

<CardGroup cols={2}>
  <Card title="Quick Start" icon="rocket" href="/getting-started/quickstart">
    Get up and running in 5 minutes with a real example.
  </Card>

  <Card title="Architecture" icon="building" href="/getting-started/architecture">
    Understand the complete Agent-CoreX architecture.
  </Card>

  <Card title="MCP Concepts" icon="book" href="/mcp/what-is-mcp">
    Learn about the Model Context Protocol and MCP servers.
  </Card>

  <Card title="Try Live" icon="sparkles" href="https://www.agent-corex.com">
    Start using Agent-CoreX right now — free tier available.
  </Card>
</CardGroup>

***

## FAQ

<Accordion title="Is Agent-CoreX only for AI agents?">
  No. While designed for AI agents, Agent-CoreX can be used by any application that needs dynamic tool discovery and execution via MCP servers.
</Accordion>

<Accordion title="Does Agent-CoreX support custom tools?">
  Yes. You can create custom MCP servers or wrap existing tools as MCP servers and integrate them with Agent-CoreX.
</Accordion>

<Accordion title="What MCP servers does Agent-CoreX support?">
  Agent-CoreX supports 100+ MCP servers in the ecosystem, including GitHub, Jira, Slack, Terraform, Docker, and more. Check our MCP server guide for the complete list.
</Accordion>

<Accordion title="Can I use Agent-CoreX in production?">
  Absolutely. Agent-CoreX is production-ready with enterprise-grade security, rate limiting, and monitoring.
</Accordion>

<Accordion title="Do I need to use Claude Code?">
  No. While Agent-CoreX works great with Claude Code, it integrates with any application via REST API.
</Accordion>

***

**Ready to build intelligent, adaptive agents?** Start with our [Quickstart Guide](/getting-started/quickstart) or explore [real-world use cases](/use-cases/github-automation).