Unstructured Technologies logo


Unstructured Technologies Products and Solutions

Products

1) Platform 

For users who require reliable, customizable, scalable, open-architecture pipelines, our Platform provides the ability to transform, chunk, enrich, embed, and manage an AI-ready data layer. Designed to support everything from basic RAG workflows to enterprise-scale, multimodal AI and agentic applications, the Platform scales across use cases, from fast, non-GPU pipelines for developers to highly configurable solutions for large organizations handling diverse data modalities. The Platform also includes plug-in capabilities to enhance workflows with additional enrichment, metadata extraction, and processing customizations.

Features

AI-Ready Data: 

  • Connect, transform, enrich, embed, and provide additional plug-in enhancements to build and manage your AI-ready data layer across RAG applications and other AI/ML workflows.
  • Available via API or with a low-code GUI. 
  • Transform your data using three different strategies based on the complexity of the data; access new transformation strategies as they become available.

Reliable

  • Support for 60+ file types and expanding support for additional formats and modalities.
  • Configure 35+ connectors to retrieve data wherever it lives
  • Analytics dashboard for usage insights.
  • Schedule when and how you retrieve, preprocess, and stage your data.

Extensible

  • Built on Unstructured's open-source technology, which has been downloaded millions of times.
  • Developed with a modular open systems approach to be customizable with modules and avoid vendor lock-in.
  • Compatible with any embedding model, vector database, and LLM framework.

Performant

  • Next-generation vision transformer for images, PDF, and table extraction
  • Enhanced models for table extraction, document hierarchy and element classification
  • Reduce processing time by transforming as many documents as needed simultaneously.
  • Access to ongoing feature and performance improvements

Infrastructure

  • Deployable via SaaS, in government cloud, on-premises, or hybrid.
  • IL-5 ready with plans for IL-6 availability.
  • CPU and GPU configurations

2) API Plus

Building on the success of our open-source API, which is widely used across the Federal Government, we have made available an updated version with similar architecture and drastically improved transformation quality leveraging cutting edge vision transformers in addition to traditional and computationally efficient techniques. 

Features

  • Local file ingestion for fast prototyping
  • Next-generation vision transformer for images, PDF, and table extraction
  • Enhanced models for table extraction, document hierarchy and element classification
  • Supports 50+ languages
  • Preprocess one document at a time
  • Compatible with any vector database and LLM framework
  • Multiple pipelines available that optimize speed, accuracy and ease of use
Back to Top