xby-ocr-pass

When processing immigration, visa-filing, or identity-verification workflows, the key fields on HK/Macau permits and Taiwan permits—permit number, name, gender, date of birth, validity period, and place of issue—often have to be typed in by hand, which is slow and error-prone. xby-ocr-pass delegates this task to model inference: supply a single photo of the permit and the plugin extracts every field above, parses the MRZ machine-readable zone at the bottom, and returns structured text.

The plugin offers three image-input channels—file URL, Base64-encoded string, and local file path—so it fits web integrations, direct API calls, and local scripts alike. All recognition results come back in a single function call with no extra post-processing. The API key can be set with one chat command and is persisted automatically, surviving restarts without reconfiguration.

It suits developers building batch permit-image pipelines, visa service centres, immigration data-entry teams, and any individual who needs to turn scanned or photographed documents into structured data quickly. If you already run a DeepSeek Harness workflow on document images, xby-ocr-pass slots in as a ready-to-use OCR node.