What's New

User-visible improvements to the platform. Most recent first.

February 23, 2026 Newly Shipped

Document Intelligence — See Everything Inside Your Documents

  • ▸ Images in documents are now visible. Embedded images in PDFs and Word documents appear in the document viewer, inline with the surrounding text, right where the document placed them. No downloading originals to see what's inside.
  • ▸ Image files open directly in the viewer. Upload a JPEG, PNG, TIFF, or other image format and it opens in the document viewer. When the document is primarily visual with minimal text, the image view opens automatically.
  • ▸ Discover mode. A new way to work through a document set — start from the entities that appear in your documents rather than a pre-formed keyword. Filter by entity type (person, organization, email, location), date range, and file type. Facet counts show document volume per entity. Every result opens inline in the document viewer.
  • ▸ Graph Explorer: evidence behind every connection. Click any edge between two entities and see the source documents that place them together. The connection is no longer just a line — it links to the evidence.
  • ▸ Graph auto-clustering. Large graphs with dense entity networks now auto-cluster, preventing the hairball problem. Clusters can be expanded individually to explore sub-networks.
  • ▸ In-document search navigation. Once you open a document, search continues to guide you — highlighted matches, previous/next navigation, jumping directly to the relevant passage. Works in both Search and Discover views.
  • ▸ Phrase search hint. When you type two or more words with no operators, a one-click hint offers to convert the query to a phrase search — finding documents where those words appear together, in order.
February 6, 2026

Epstein Files — Public Search Now Live

  • ▸ All ingested documents from the DOJ's public Epstein release are searchable at epstein-files.ingestigate.com — free, no credit card required.
  • ▸ Full entity extraction across 849,000+ documents. Browse by person, organization, email, phone number, and more.
  • ▸ Relationship graph available for all extracted entities.
2025

Core Platform

  • ▸ Named Entity Recognition (NER). Automatic extraction of people, organizations, email addresses, phone numbers, cryptocurrency addresses, usernames, and more — running entirely on your infrastructure, no external API calls.
  • ▸ Graph Explorer. Visual relationship network built from entity co-occurrence across your documents. Force-directed layout, edge weights by document frequency.
  • ▸ Owner-controlled access. Investigation owners decide who can see their data. Other users in your organization can see that an investigation exists, but cannot read its contents without explicit access.
  • ▸ Sub-millisecond search. 0.2ms average query response across millions of documents. Full-text, phrase, proximity, and include/exclude operators.
  • ▸ Structured data search. Search inside Excel columns, JSON, CSV, Parquet, and ORC files directly — no export required.
  • ▸ On-premise and air-gapped deployment. Full platform in a containerized deployment for environments where data cannot leave your network.

Questions about a specific change? Contact us or read the blog for deeper context on major updates.