Skip to main content

Workflows and Limits

Common questions about document processing workflows, throughput limits, and queue behavior.

Q: How long does document processing take?​

Most documents complete OCR extraction within 30 seconds to a few minutes. Processing time depends on file size, document complexity, and current queue depth. Multi-document PDFs take longer since each form is processed separately after the separation engine splits the bundle.

You can monitor progress in real time on the upload page. Status updates move from Uploaded to Processing to OCR Complete as documents advance through the pipeline.

Q: What happens if I upload a very large PDF bundle?​

Multi-document PDFs are split automatically into individual forms. A 50-page packet containing 12 tax forms creates 12 separate review cases. The separation engine identifies form boundaries using layout analysis.

For best results with large bundles, ensure each form follows a standard IRS layout. Non-standard pages or non-tax documents may reduce separation accuracy. If you encounter issues, try uploading forms individually.

Q: Is there a limit on how many documents I can upload at once?​

There is no hard limit on simultaneous uploads. The upload queue accepts multiple files and processes them sequentially through the OCR pipeline. Large batches spread processing across the available worker capacity.

If you are processing thousands of documents, contact support to discuss volume expectations and potential queue prioritization options.

Q: What if OCR confidence is low on a document?​

Low confidence scores trigger the document for manual review in the Review Hub. You can correct any extracted field before exporting. Low confidence typically results from poor scan quality, handwriting, or non-standard form layouts.

All fields remain editable regardless of confidence level. The confidence score helps you prioritize which documents need closer attention.

Q: Can I reprocess a document if something goes wrong?​

Yes. If a document fails during processing or you need to restart extraction, you can trigger reprocessing from the document actions menu. This sends the document back through the OCR pipeline.

Reprocessing is useful when the original file had quality issues that have since been corrected, or when you want fresh extraction after a system update.

Q: How does the confidence-based approval workflow work?​

Documents that pass all validation rules with high confidence scores may be pre-approved, meaning they require less manual verification. The system uses confidence thresholds and form-specific validation rules to determine which cases can skip detailed review.

You can always review pre-approved documents if you prefer. The confidence system reduces workload without removing your control over final data quality.

Q: What status values can a document have?​

Documents move through these statuses:

  • Uploaded - File received, awaiting processing
  • Processing - OCR extraction in progress
  • OCR Complete - Extraction finished, pending review
  • Needs Review - Ready for human verification
  • Approved - Verified and ready for export
  • Exported - Data exported to file package

Error states include Failed and Unidentified for documents that could not be processed or classified.

Q: How do I handle documents that cannot be classified?​

Documents that the system cannot identify appear with an Unidentified status. You can manually assign the correct form type in the Review Hub. If the document is not a supported form type, you may need to process it outside TidalForms.

Check that the document is legible, uses a standard layout, and is one of the supported form types. State variants and heavily modified forms may not be recognized.