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

# PDF

> Extract text, metadata, and page content from PDF files.

# PDF

Extract text, metadata, and page content from PDF files. Accepts file paths, URLs, and base64-encoded data.

<Info>
  Requires the `pdf-parse` peer dependency.
</Info>

***

## Quick Start

```typescript theme={null}
import { Agent, openai, PdfToolkit } from "@agentium/core";

const pdf = new PdfToolkit();

const agent = new Agent({
  name: "document-analyst",
  model: openai("gpt-4o"),
  instructions: "Extract and summarize content from PDF documents.",
  tools: [...pdf.getTools()],
});

const result = await agent.run("Extract the text from ./report.pdf and summarize the key findings.");
```

***

## Config

<ParamField body="maxLength" type="number" default="50000">
  Maximum characters to return per extraction. Longer text is truncated.
</ParamField>

***

## Tools

| Tool                | Description                                                            |
| ------------------- | ---------------------------------------------------------------------- |
| `pdf_extract_text`  | Extract all text from a PDF. Accepts a file path, URL, or base64 data. |
| `pdf_get_metadata`  | Get PDF metadata — title, author, page count, creation date.           |
| `pdf_extract_pages` | Extract text from specific pages (1-indexed).                          |

***

## Peer Dependency

```bash theme={null}
npm install pdf-parse
```

***

## Input Sources

The `source` parameter accepts three formats:

* **File path**: `/path/to/document.pdf`
* **URL**: `https://example.com/report.pdf`
* **Base64**: Raw base64-encoded PDF data

```typescript theme={null}
// From file
await agent.run("Extract text from /tmp/invoice.pdf");

// From URL
await agent.run("Extract text from https://example.com/whitepaper.pdf");
```