Docs · 1 credit
DOCX extract
Give agents a local DOCX extract path that matches how you already handle PDFs — same host, same auth, chosen tools only. Body text is what=text.
What it does
Call this tool (id: docx_extract) to extract from a Word .docx: body text, comments, headers/footers, hyperlinks, images, tables, heading outline.
Example use cases
- HR or legal agents reading .docx uploads.
- Migration bots that normalize Word content to plain text.
- Mixed document inboxes (PDF + DOCX) behind one MCP endpoint.
How it works in Toolporten
Compose docx_extract (what=text for body text) with pdf_text_extract when the agent should handle both formats. Former child ids such as docx_text_extract are disabled; ToolAliasMap redirects them. Toolporten’s model is the slim allowlist — not a universal document mega-server.
MCP call example
{
"name": "docx_extract",
"arguments": {
"what": "text",
"docx_url": "<docx-url>"
}
}