The Document Parser node extracts text from documents. Point it at a URL or upload files, and it returns the document’s text as a single block you can pass to later nodes.
What it does
- Parses documents such as PDFs and office files into clean text.
- Reads from a URL or from files you upload.
- Handles tables, columns, and specific target pages through parsing settings.
Configuration reference
| Field | Required | What it does |
|---|---|---|
| Input type | Yes | Link to parse a document by URL, or Upload to parse uploaded files. Defaults to Link. |
| Document link | When Link | The document URL. Supports variable tags. |
| Uploaded files | When Upload | The files to parse. |
| Parser settings | No | Controls for language, parsing instructions, skipping diagonal text, keeping columns unrolled, and target pages. |
Local or Cloud, and cost
Document Parser runs in both Local and Cloud modes and costs 0 credits.
Using the result in later nodes
The parser produces the extracted text as one block, referenced with [Input from <this-node-id>]. It does not split the text into named fields. If you need named fields from that text, add an AI Extract Information node after it.
Worked example: read a PDF report
- Add a Document Parser node, set input type to Link, and set the document link to
https://example.com/report.pdf. - Add an AI Prompt node that summarizes
[Input from parser-1].
Related
Was this helpful?
Contact us