---
title: Using LlamaIndex with AI Agents | Developer Documentation
description: A map of the MCP servers, agent skills, plugins, and workflow nodes you can use to give an AI agent LlamaIndex capabilities.
---

There are multiple ways to give an AI agent our capabilities: connect it to an **MCP server**, install an **agent skill or plugin**, or drop a **workflow node** into a no-code pipeline.

## Pick your path

| What you want                                           | Use this                                                       |
| ------------------------------------------------------- | -------------------------------------------------------------- |
| My coding agent should parse and extract from documents | [Agent skills and plugins](#agent-skills-and-plugins)          |
| I want my agent to use the LlamaParse Platform remotely | [LlamaParse Platform MCP](/llamaparse/for-agents/mcp/index.md) |
| My agent should be able to search these docs            | [Documentation MCP](/for-agents/mcp/index.md)                  |
| I’m building a no-code document workflow                | [n8n node](/llamaparse/for-agents/n8n/index.md)                |
| I’m reading these docs programmatically                 | [Programmatic docs access](#programmatic-docs-access)          |

## MCP servers

We host two [Model Context Protocol](https://modelcontextprotocol.io/) servers.

### LlamaParse Platform

`https://mcp.llamaindex.ai/mcp` is the endpoint that exposes the platform’s document processing — Parse, Classify, Extract, Split, and Index — as tools any MCP client can call. Authentication is OAuth, so there’s no API key to manage. Product-scoped servers (for example `https://mcp.llamaindex.ai/extract/{configId}/mcp`) narrow the tool list when you want tighter scoping or more parallelism.

[Learn more about LlamaParse MCP setup](/llamaparse/for-agents/mcp/index.md)

### Documentation search

`https://developers.llamaindex.ai/mcp` gives an agent `search_docs`, `grep_docs`, and `read_doc` over this documentation site. Useful when you want an agent writing LlamaIndex code to look things up instead of guessing.

[Learn more about documentation MCP setup](/for-agents/mcp/index.md)

## Agent skills and plugins

**Skills** are self-contained capabilities a coding agent loads on demand or through direct slash-commands. Two are available, both from [`run-llama/llamaparse-agent-skills`](https://github.com/run-llama/llamaparse-agent-skills):

| Skill        | What it does                                                                               | Requires              |
| ------------ | ------------------------------------------------------------------------------------------ | --------------------- |
| `liteparse`  | Local-first parsing of PDFs, Office docs, and images. Fast, runs entirely on your machine. | Node 18+, no API key  |
| `llamaparse` | Cloud parsing built for complex documents with charts and tables.                          | `LLAMA_CLOUD_API_KEY` |

- [Learn more about the LiteParse agent skill](/liteparse/guides/agent-skill/index.md)
- [Learn more about LlamaParse skills and plugins](/llamaparse/for-agents/skills/index.md)

## Agent Plugins

**Plugins** bundle the skills for Claude Code and Codex so you can install them from inside the agent. Three are available: `liteparse`, `llamaparse`, and `llamaparse-mcp` — the last bundling a skill together with the LlamaParse Platform MCP.

[Learn more about installing plugins](/llamaparse/for-agents/skills/index.md)

## Workflow integrations

The **n8n community node** brings the LlamaParse Platform into n8n workflows, so you can build document pipelines without writing code.

[Learn more about the LlamaParse Platform n8n node](/llamaparse/for-agents/n8n/index.md)

## Programmatic docs access

This documentation site is built to be read by agents, not just people:

- **`llms.txt`** — an overview of the whole site at [`developers.llamaindex.ai/llms.txt`](https://developers.llamaindex.ai/llms.txt)
- **Raw Markdown** — append `index.md` to any page URL, e.g. [`/llamaparse/parse/getting_started/index.md`](https://developers.llamaindex.ai/llamaparse/parse/getting_started/index.md)
- **REST API** — `/api/search` (BM25 full-text), `/api/grep` (regex), `/api/read` (fetch a page), and `/api/list` (browse the tree). Parameters are documented in `llms.txt`.

If your client supports MCP, the [documentation MCP server](/for-agents/mcp/index.md) wraps the same functionality and is usually preffered.