Docs · 1 credit
PDF text extract
Pull per-page and combined text from PDFs so agents can summarize, cite, or route documents without a separate parse service.
What it does
Call this tool (id: pdf_text_extract) to extract text layers from a PDF (base64). Returns per-page text and combined text.
Example use cases
- Ingesting invoices and contracts into an agent loop.
- Support bots that read attached PDFs and answer from the text layer.
- Pipelines that extract text, then chunk or redact with other toolbox tools.
How it works in Toolporten
Allowlist pdf_text_extract (and optionally docx_extract with what=text) for document-heavy agents. They call your toolbox MCP URL with an API key; you keep the surface small and credit-metered.
MCP call example
{
"name": "pdf_text_extract",
"arguments": {
"pdf": "<base64-pdf>",
"maxPages": 10
}
}