Skip to main content

Overview

Hyland’s Knowledge Enrichment APIs convert raw, unstructured content into structured and enriched data that is optimized for AI, automation, analytics, and solution development. These APIs extract key entities, apply metadata, and enrich content with contextual links, ensuring data is machine-readable and actionable from the start.

Some of the key capabilities of Knowledge Enrichment APIs include the following:

CapabilityDescription
Entity ExtractionIdentify and extract meaningful entities from documents.
Metadata TaggingAutomatically assign metadata for improved organization and retrieval.
Contextual LinkingEnhance content by linking it to relevant contextual information.
Streamlined Data PreparationAutomate extraction and enrichment tasks.
Accelerated AI DeploymentDeliver machine-ready data faster.
Reduced Manual EffortMinimize time spent on cleaning and formatting.
Data Integrity and ComplianceEnsure consistent handling of sensitive or regulated content.

The Hyland Knowledge Enrichment APIs include the following core components:

How the core APIs work together​

The following diagram provides a visual representation of how the core APIs work together. The Data Curation API (DC API) and Context API form a sequential preprocessing pipeline. The DC API runs first and produces clean, normalized text. The Context API runs second and adds semantic meaning on top of that text.

A consumer can use the DC API alone (for text extraction and chunking) or chain both APIs. The combined output — structured JSON containing text, metadata, chunk positions, and optionally embeddings — is consumed by RAG pipelines, AI agents, and search indexes.

Data Curation API​

The Hyland Data Curation API is designed to transform raw, unstructured content into structured data suitable for AI and machine learning applications. As part of Hyland's Content Innovation Cloud, this API streamlines the extraction, enrichment, and structuring of content from a wide range of file types, including documents, images, and audio files.

By automating key data preparation steps, such as text extraction, PII redaction, and content chunking, the DC API provides the following benefits:

  • Reduces manual effort.
  • Accelerates AI model readiness.
  • Enhances the accuracy of downstream applications, such as search engines, recommendation systems, and document-understanding platforms.

Built for scalability, flexibility, and resilience, the DC API ensures that organizations can efficiently process increasing volumes of unstructured data while maintaining compliance and data privacy.

The DC API accepts raw files and runs them through a processing pipeline:

  1. Format identification — Determines the file type.
  2. Text extraction — Uses Document Filters for digital documents (text extracted directly, not OCR'd), OCR-based extraction for image-based content, and speech transcription for audio/video.
  3. Normalization — Uses consistent Unicode, configurable quote and dash normalization.
  4. PII detection and masking (optional) — Detects and redacts sensitive data before storage.
  5. Semantic chunking — Uses configurable chunk size and overlap with positional metadata.
  6. Embedding generation (optional); Uses vector per chunk.

Output is stored in the Content Lake with per-environment isolation. Processing is asynchronous.

Key advantages​

Consider the following advantages of using the Data Curation API:

  • Uniform input format — Markdown output for further processing, independent of original file type.
  • Rich format support — Handles PDFs, Office files, emails, archives, and 600+ other formats.
  • Centralized parsing logic — Delegates document parsing to a centralized, audited service.
  • Token-based access control — Secure access without internal credential management.

→ Data Curation API reference · Curating documents guide

Context API​

The Context API is designed to infer new metadata for ingested content, making it more structured, searchable, and enriched. Its primary goal is to analyze unstructured or semi-structured data, such as documents or images, and generate meaningful metadata that improves organization, discoverability, and usability of the content.

The API is read-only; it returns structured metadata without modifying source content. Actions are selective and policy-driven, and so not every document requires every action. Processing is asynchronous.

Key functions​

The key functions of the Context API include the following:

FunctionDescriptionExample
Metadata Extraction and EnrichmentAutomatically generates metadata fields, such as document type, keywords, topics, and entities. This makes retrieval easier.Extracting "Contract Type: NDA" from a legal document.
Content Categorization and StructuringAssigns content to relevant categories based on AI-driven analysis.Identifying that an uploaded file is a "Medical Report" vs. "Invoice".
Semantic Understanding and Contextual AnalysisUses NLP and AI models to understand content meaning, detecting sentiment or subject matter or generating a summary.Summarizing key points from a research paper.
Multi-Format SupportProcesses both text documents and images.Extracting actionable insights and deadlines from a meeting transcript.
Integration with Search and Retrieval SystemsEnhances search capabilities by adding structured metadata.Enables users to find documents by searching for extracted topics.
Automated Tagging and ClassificationGenerates labels to improve content discovery.Tagging an image with "Office Environment" or "Legal Document".

Interaction flow​

The following diagram illustrates the usual order in which the Context API calls should be made to get the expected results. The typical, structured sequence involves file upload, processing, and result retrieval.

Next steps​