LDXhubInstall

Summary

Upload a PDF or image from the selected binary field to LDXhub, submit it to a vision AI model with extraction instructions and an example JSON structure, then poll until the analysis job completes and download its result.

Inputs

• Input Binary Field (required) — Binary property containing the PDF or image; defaults to data.
• Model Name or ID (required) — Vision AI model to use. Available models are loaded from LDXhub; the output format choices depend on the selected model.
• System Prompt (required) — Instructions describing what structured data to extract and how.
• Example Output (required) — JSON object with example values showing the desired output structure.
• Output Format Name or ID (required) — Format for the downloaded result file, selected from formats supported by the chosen model.
• Polling Settings — Optional job polling controls: maximum attempts (default 180) and server wait seconds per poll (default 10; allowed range 1–10).

Output shape

One output item per input item, with JSON metadata and the downloaded result in binary data.

JSON contains job_id, model, status, output_file_id, and usage. The result binary uses the same property name as the input binary field. A completed job without an output file ID, a failed job, or exhausting polling attempts raises an error; when Continue On Fail is enabled, the affected item instead carries its original JSON and an error.

Links

• n8n community nodes installation guide