Get document text
doc://host/file/query/text
{
"path": "/docs/umowa.pdf"
}
Documents connector
Document text / OCR
doc:// turns document text extraction into a first-class URI instead of ad-hoc shell + pdftotext. doc://host/file/query/text returns the text of a PDF or image: a text-layer PDF goes through poppler pdftotext; a scan/image falls back to OCR, rasterising PDF pages (pdftoppm/pdfimages/magick) and running the first available engine — tesseract, then RapidOCR (the ONNX engine reused from wronai/img2nl, no system binary needed), then an Ollama vision model (reused from wronai/ocr). doc://host/file/query/ocr forces OCR; doc://host/engine/query/list reports which engines are present. Useful for classifying invoices and any scanned document over the mesh.
Co robi
- Extract the text layer of a PDF inside a flow.
- OCR a scanned invoice or image that has no text layer.
- Check which OCR engines are available on a node.
Schematy URI
doc://
Wymagania
ArrayArrayArray
Trasy
Przykłady
Force OCR on a scan
doc://host/file/query/ocr
{
"path": "/scans/faktura.png"
}
Komenda instalacji
curl -fsSL 'https://connect.ifuri.com/install?connectors=doc' | bash
Wpis rejestru
{
"id": "doc",
"name": "Document text / OCR",
"status": "available",
"uriSchemes": [
"doc"
],
"routes": [
"doc://host/file/query/text",
"doc://host/file/query/ocr",
"doc://host/engine/query/list"
],
"install": {
"mode": "urirun-extra",
"pipSpec": "urirun-connector-doc @ git+https://github.com/if-uri/urirun-connector-doc.git@v0.1.0"
},
"hubUrl": "https://connect.ifuri.com/connectors/doc"
}