In July, we focused on solidifying the foundation Hypha's Synthesis and Cortex product. We believe that by investing heavily in quality and fidelity of the extraction layer, we'll reap the benefit with more accurate analysis (Cortex) and workflow automation (Flow) downstream.
New feature highlights
AMA (Ask Me Anything):
Ask your question in plain language—Hypha does the rest. Our platform translates everyday queries into pinpoint data results, revealing predictive insights you can use right now. No SQL. No scripts. Just answers you can act on.
Audit trail for both human and AI activities:
Easily track and compare every version of an AI‑generated answer, then drill into each human edit and approval at the entity or attribute level to validate accuracy and prove compliance in one streamlined view.
Multi-modal support with enhanced OCR:
Multi‑modal input with enhanced OCR captures text, tables, grids, and images exactly as they appear, preserving structure and detail. The result? Cleaner data, fewer corrections, and higher downstream accuracy for Cortex and Flow.
Agentic document chunking:
Hypha redefines document processing—breaking content into efficient, token-light segments while flawlessly preserving structure and context. From intricate tables to multi-level lists, every detail stays intact, powering faster, smarter, and more cost-efficient answers.
UX Improvements
This month, we have been laser focused on removing friction from user experiences. Here are examples of some QoL improvements we've made after hearing your feedback.
Added keyboard shortcuts, speed up your navigation and approval
Added ability to link attribute to specific document types, fully control the extraction process
Added ability to re-run entity extraction
Export attributes to excel, word and PDF, collaborate seamlessly with your team
Expanded support for output types beyond just text to match your formatting preferences
Currency
Percent
Number
Date
Yes/No
Get started
Book a demo today!
