# Docs

Document extraction for agents: download a public URL or pass base64, then return text/tables. PDF uses pdfplumber; empty/scanned pages fall back to poppler + Tesseract OCR with language packs (default hun+eng).

**Service id:** `docs`  
**Version:** `0.1.0`  
**Status:** `available`

## Authentication

MCP endpoint: `https://mcp.glc-rag.hu/mcp` (streamable HTTP)

**Agents (recommended):** self-register with `account_type=agent` to get an
auto-approved token — see [https://mcp.glc-rag.hu/guide/agent](https://mcp.glc-rag.hu/guide/agent).

Or register as a human on the public site, request this service in Admin,
wait for system-admin approval, then create a token.

```http
Authorization: Bearer mcp_...
```

Cursor `mcp.json` example:

```json
{
  "mcpServers": {
    "docs": {
      "url": "https://mcp.glc-rag.hu/mcp",
      "headers": {
        "Authorization": "Bearer mcp_YOUR_TOKEN"
      }
    }
  }
}
```

## Tools

### `docs_extract`

Extract text and tables from a document (pdf, docx, xlsx, txt, csv). Provide either a public http(s) URL or base64 content. For scanned PDFs, OCR runs automatically when little text is found.

**Input schema:**

```json
{
  "type": "object",
  "properties": {
    "url": {
      "type": "string",
      "description": "Public http(s) URL of the document"
    },
    "content_base64": {
      "type": "string",
      "description": "Raw document bytes as base64 (or data: URL)"
    },
    "filename": {
      "type": "string",
      "description": "Optional filename hint for type detection (e.g. report.pdf)"
    },
    "lang": {
      "type": "string",
      "description": "OCR languages, e.g. hun+eng (default hun+eng)",
      "default": "hun+eng"
    },
    "force_ocr": {
      "type": "boolean",
      "description": "Always OCR PDF pages even if embedded text exists",
      "default": false
    },
    "max_length": {
      "type": "integer",
      "minimum": 1,
      "maximum": 1000000,
      "default": 20000
    },
    "start_index": {
      "type": "integer",
      "minimum": 0,
      "default": 0
    },
    "max_pages": {
      "type": "integer",
      "minimum": 1,
      "maximum": 200,
      "default": 50,
      "description": "Max PDF pages to process"
    }
  },
  "additionalProperties": false
}
```

**Examples:**

```json
{
  "url": "https://www.w3.org/WAI/ER/tests/xhtml/testfiles/resources/pdf/dummy.pdf"
}
```

```json
{
  "url": "https://example.com/scan.pdf",
  "lang": "hun+eng",
  "force_ocr": true
}
```

### `docs_status`

Docs service health, supported formats, and installed OCR languages.

**Input schema:**

```json
{
  "type": "object",
  "properties": {},
  "additionalProperties": false
}
```

**Examples:**

```json
{}
```

## Usage notes

Call `docs_extract` with `url` or `content_base64`. Set `lang` to `hun`, `eng`, `deu`, or combinations like `hun+eng`. Use `force_ocr: true` for scanned PDFs. Paginate with `max_length` / `start_index`.

## Errors / limits

Unsupported type, SSRF-blocked URL, oversized body, missing OCR packs, or extract failure → `{error, is_error}`. Invalid API key → platform 401.

## Agent discovery

- Agent registration: `https://mcp.glc-rag.hu/guide/agent`
- Markdown: `https://mcp.glc-rag.hu/guide/docs.md`
- Index: `https://mcp.glc-rag.hu/llms.txt`
- MCP resource: `docs://docs`