DeepReadInstall

Summary

Submit each input document to DeepRead for OCR extraction, then poll until processing completes or the timeout is reached. The result includes the service’s job response, with an optional searchable PDF binary when that artifact is returned.

Inputs

• Input Binary Field (required) — Binary property containing the document; defaults to data.
• Engine — Processing engine sent as the pipeline. Plan Default omits this field; Extract and Deep Extract select their respective engines.
• Searchable PDF — When enabled, requests a searchable PDF and returns it as binary property searchablePdf if DeepRead provides the artifact. Defaults to off; the Extract engine does not support this feature.
• Timeout (Seconds) — Maximum time to wait for job completion; defaults to 300 seconds.
• Options: Idempotency Key — Optional request key used by DeepRead to reuse a job on retry; the service rejects a different request using the same key.
• Options: Incognito — Requests single-use-key encryption and crypto-shredding after completion. Incognito jobs do not provide previews or field locations and cannot be combined with Searchable PDF.
• Options: Preview — Requests stored page images, a public preview link, and extracted-value locations. Defaults to off; links are accessible to anyone who has them. Always off for incognito jobs.
• Options: Retention (Days) — Requests deletion of the document, preview artifacts, and extracted results after the specified 1–365 days; defaults to 30.

Output shape

One output item per successfully processed input, containing the DeepRead job response in JSON and, when requested and available, a searchable PDF in binary property searchablePdf.

The node polls the submitted job until completion, failure, or timeout. With Continue On Fail enabled, errors are returned as JSON { "error": "..." } items paired with their input.

Links

• n8n community nodes installation guide