Configuration

The configuration area in MyndDMS centralizes all system-level behavior into a structured set of controls. It defines how documents are processed, how text is extracted, how automation features behave, and how optional intelligence capabilities are integrated into the platform.

These settings influence the entire document lifecycle. They affect ingestion rules, OCR processing quality, barcode-based automation, AI-assisted features, and output generation parameters. Adjustments made here directly impact accuracy, performance, and functional capabilities across the system.

The interface also includes options that control application presentation, such as branding elements and display identifiers, ensuring that the deployment can be aligned with organizational requirements.

In essence, this section represents the operational configuration layer of myndDMS. It determines how the system processes content, applies automation, integrates advanced features, and presents itself within its environment.

Application Branding

Application Title

This parameter controls the application title displayed in the browser tab and main header. It is typically used in white-label deployments where the system name must reflect the organization’s branding rather than the default product name. The value is a plain text string.

OCR Configuration

The following parameters define how documents are processed during the Optical Character Recognition (OCR) phase. These settings influence text extraction accuracy, archival compliance, and processing performance.

Output Type

This setting determines the format of the PDF generated after OCR processing. It controls whether the output is a standard searchable PDF or an archival-compliant format such as PDF/A. The selected format may impact long-term document preservation requirements and regulatory compliance standards.

Language

This parameter specifies the language model used during text recognition. It directly affects OCR accuracy, especially for multilingual documents. Multiple languages can be defined when necessary to improve recognition results in mixed-language content.

Pages

This option limits the number of pages processed by the OCR engine. When configured, only the first defined number of pages are analyzed. If left unset, all pages are processed. This setting can reduce processing time for large documents where full OCR is not required.

Mode

This parameter defines how OCR is applied to incoming documents. It determines whether the system skips pages that already contain text, forces OCR on all pages regardless of existing content, or applies conditional processing. The selected mode affects both processing time and output consistency.

Skip Archive File

This setting controls whether the original document file is preserved alongside the OCR-processed version. Retaining the original file ensures traceability and archival integrity, while skipping it reduces storage consumption.

Image DPI

This parameter defines the resolution used when image metadata does not specify DPI. Proper DPI configuration ensures consistent rendering and reliable OCR accuracy, particularly for scanned documents of varying quality.

Clean

This setting enables image cleanup prior to OCR processing. When activated, the system attempts to remove visual noise and improve clarity, which can significantly enhance recognition accuracy in low-quality scans.

DESKEW

This option enables automatic correction of slight page misalignment. It improves OCR results by ensuring text is properly aligned before recognition.

Rotate Pages

This parameter activates automatic rotation correction for pages detected as incorrectly oriented. It ensures text is upright prior to OCR processing, improving extraction accuracy.

Rotate Pages Threshold

This setting defines the sensitivity level used to detect incorrectly rotated pages. Lower values increase the likelihood of automatic correction, while higher values make rotation detection more conservative.

Max Image Pixels

This parameter sets a maximum pixel limit for images processed during OCR. It acts as a safeguard against excessive memory usage and protects the system from processing unusually large or potentially malicious files.

Color Conversion Strategy

This setting determines how color information is handled when generating the final PDF output. Depending on the chosen strategy, the system may preserve original colors or convert documents to grayscale or alternative color spaces for archival consistency.

OCR Arguments

This parameter allows advanced configuration by passing additional arguments directly to the OCR engine. It enables fine-grained control beyond default settings and is intended for experienced administrators who require specialized OCR behavior.

Barcode Processing

Enable Barcodes

This setting enables the barcode engine during document consumption. When activated, incoming files are analyzed for supported barcode formats, and all related barcode features become available.

Enable TIFF Support

Extends barcode detection to multi-page TIFF documents. This is required when scanned documents are stored as .tif or .tiff files rather than PDFs.

In workflows where scanners output image-based formats, enabling this option ensures consistent barcode processing across file types.

Barcode String

Defines the barcode value that acts as a document split trigger; when this specific barcode is detected during processing, MyndDMS creates a new document starting from that page.

Retain Split Pages

Controls whether pages containing the split barcode are preserved after document separation.

  • Enabled: Separator pages remain in the resulting documents.

  • Disabled: Separator pages are removed.

This behavior can be adjusted depending on archival or compliance requirements.

Enable ASN

Enables Automatic Serial Number (ASN) barcode recognition.

When active, barcodes matching ASN rules are interpreted as structured identifiers and can be used for document identification or automation workflows.

ASN Prefix

Defines the required prefix for valid ASN barcodes.

Only barcodes beginning with this prefix are processed as ASN values.
This prevents unintended barcode matches and enforces structured formatting.

Upscale

When enabled, documents are upscaled before barcode detection.

This improves recognition accuracy for:

  • Small barcodes

  • Low-resolution scans

  • Poor-quality images

Trade-off: Increased CPU usage and processing time.

DPI

Specifies the resolution used when converting documents for barcode scanning.

Higher DPI values:

  • Improve detection reliability

  • Increase memory and processing requirements

The optimal value depends on document quality and system resources.

Max Pages

Limits the number of pages analyzed for barcode detection.

  • 0 → No limit (all pages scanned)

  • Any positive integer → Only that number of pages is processed

Useful for performance optimization in large documents.

Enable Tag Detection

Enables automatic tag assignment based on detected barcode values.

When activated, barcode-to-tag mappings are evaluated during ingestion, and matching tags are applied automatically.

Tag Mapping

Defines the relationship between barcode values and tags.

When a detected barcode matches a configured value, the corresponding tag is assigned to the document.

Multiple mappings can be configured to support structured classification workflows.

Split on Tag Barcodes

Determines whether tag barcodes also trigger document splitting.

  • Enabled: The barcode assigns a tag and splits the document.

  • Disabled: The barcode assigns a tag only.

This allows flexible behavior depending on operational requirements.

AI Settings

AI Enabled

Controls whether AI features are active within the system. When enabled, the system can perform tasks such as automatic metadata generation, intelligent tagging, natural-language retrieval, or document summarization. Be aware that enabling remote AI processing may have privacy implications, as document content could be transmitted to a third-party service.

LLM Embedding Backend

Specifies the service used to generate semantic embeddings — numeric vector representations of text used for similarity search and retrieval-augmented generation (RAG). The chosen backend determines how query and document text are embedded for search and contextual understanding. Common options include hosted providers (like Hugging Face) or local hosts (like Ollama), depending on performance and privacy requirements.

LLM Embedding Model

Defines the specific embedding model used by the embedding backend. The model transforms raw text into vectors that capture semantic relationships between documents and queries. The choice of model affects retrieval quality, memory usage, and speed. For example, larger models typically produce higher-quality vectors but require more resources.

LLM Backend

Specifies the large language model (LLM) provider or host for general AI tasks beyond embeddings. This setting dictates where the system sends text for generation, summarization, or question-answering. It can point to a cloud API, a self-hosted LLM server, or a local inference engine.

LLM Model

Identifies the specific language model to use for generative tasks. This can be a model hosted on a cloud provider (e.g., an OpenAI or Hugging Face model) or a locally hosted model (e.g., through Ollama). The selected model determines capabilities such as response coherence, multilingual support, and resource requirements.

LLM API Key

Stores the authentication key required to access a remote LLM provider. This key is used by the backend service to authenticate requests to the cloud API. It is treated as a secret and typically masked in UI displays. Without a valid API key, remote AI services will not function.

LLM Endpoint

Defines the network endpoint (URL) where AI requests are sent. For remote providers, this is typically an HTTP API base URL. For self-hosted inference, it may point to a local server address. The endpoint is used by the system to dispatch requests for text generation, embeddings, or model inference.