---
title: "LLMind quickstart — install and enrich your first file in 5 minutes | LLMind"
description: "Get started with LLMind in 5 minutes. Install from source, enrich a document, verify the semantic layer, and use enriched files with Claude and other AI tools."
url: https://llmind.org/docs/quickstart/
source_format: html
---
# LLMind quickstart

Published 2026-04-22 · 5 min read

Get started with LLMind in 5 minutes. You'll install the tool, enrich your first document, verify the semantic layer was embedded, and learn how to use enriched files with Claude and other AI tools.

## Step 1: Install LLMind

LLMind ships as a Python package on PyPI. Install it with `pipx` (recommended) or `uv` — both put the CLI on your `PATH` in an isolated environment without touching your system Python.

Don't have `pipx` yet? On macOS run `brew install pipx && pipx ensurepath`; on Debian/Ubuntu run `sudo apt install pipx && pipx ensurepath`. Open a new terminal afterward. Full per-platform instructions in the [install guide](https://llmind.org/docs/install/#install-pipx).

```
pipx install 'llmind-cli[all]'
llmind --version
```

Requires Python 3.11+. Pick a vision provider with the install extra: `[anthropic]` for Claude, `[openai]` for GPT-4o, `[gemini]` for Gemini, or `[all]` for every provider. PDF rendering also needs `poppler-utils` on your system (`brew install poppler` on macOS, `apt install poppler-utils` on Debian/Ubuntu). Works on macOS and Linux.

## Step 2: Enrich a document

Once installed, you can enrich any document (PDF, image, video, etc.):

```
llmind enrich document.pdf
```

LLMind analyzes the document, extracts structure and entities, and writes a semantic layer inside the file as XMP metadata. The file is modified in-place; a backup is recommended for irreplaceable originals.

## Step 3: Verify the layer

To see what was embedded, run:

```
llmind inspect document.pdf
```

This shows the semantic layer: a description, extracted entities, structural summary, and any signatures (if verification is enabled). The layer is now portable—it travels with the file wherever it goes.

## Step 4: Use enriched files with Claude

To give Claude access to your enriched files, configure Claude Desktop to run the filesystem MCP server pointed at a directory. Claude will then read your enriched files and automatically parse the semantic layer alongside the raw content.

See the [Claude Desktop MCP setup guide](https://llmind.org/mcp/claude-desktop/) for step-by-step configuration and examples.

## What's next?

Now that you've enriched your first file, explore these guides:

-   [Install](https://llmind.org/docs/install/) — detailed installation options and platform-specific notes
-   [CLI reference](https://llmind.org/docs/cli/) — all commands, flags, and options
-   [What is file enrichment?](https://llmind.org/learn/what-is-file-enrichment/) — conceptual overview
-   [AI agents + file access](https://llmind.org/use-cases/ai-agents-file-access/) — patterns for using enriched files with autonomous tools
-   [Product](https://llmind.org/product/) — overview of LLMind