​
August 11, 2025

​
What’s new

Aryn DocParse now supports HTML as output, simply specify html in the output_format option. This is useful when converting documents for RAG and other LLM-based applications.

​
Improvements

  • We’ve improved the performance of document listing by 10x, which is especially beneficial for customers that have hundreds of thousands to millions of documents.
  • We’ve made Aryn console more interactive and responsive for large and dense tables.
​
August 4, 2025

​
What’s new

  • Aryn DocParse now supports parsing files referenced from URLs. You can pass in pre-signed URLs from Amazon S3, Google Cloud Storage, Azure Blob Storage, or elsewhere. Aryn efficiently and reliably parses the linked files without cumbersome file management on the client side. For more information, visit the Aryn documentation.
  • Aryn DocParse also now supports linking the captions associated with an image in a document. Simply add image_extraction_options: {'associate_captions': true} and you get images expanded to encompass the captions. For more information, take a look at the Image Extraction tutorial.

​
Improvements

We’ve improved the performance and overall look and feel for DocSet browsing and the DocParse UI. Try it out and let us know what you think!
​
July 28, 2025

​
What’s new

We’re proud to announce that Aryn has attained SOC 2 Type II and ISO 270001 compliance! We’ve always cared deeply about our customer’s data security, privacy, and availability. This certification validates what we’ve been doing all along — going beyond the industry’s best practices to protect our customer’s data — not once but continuously.

​
Improvements

We improved the throughput of our VLM-based document processing by more than 100%. This improves both VLM-based table extraction (table_mode = vision) and VLM-based OCR (text_mode = vision_ocr ), allowing you to process twice the number of pages in the same time budget. Try it out on your most complex docs.