Seven-person startup Unsiloed AI has raised a $7.5M seed round led by Felicis to build a computable document layer for AI agents. Its dual-stream vision models read content and layout in parallel, heat-map chunking concentrates compute on figures and charts, every extracted field links back to its page location, and the system can run air-gapped on customer servers.
Unsiloed AI
Document-intelligence startup converting PDFs, scans and spreadsheets into structured, queryable data for AI agents using dual-s
Curated Intelligence
Rohan Paul (@rohanpaul_ai)T2·Industry
62 pts
TechCrunch AIT2·Industry
62 pts
Unsiloed AI has raised a $7.5 million seed round led by Felicis, with participation from Y Combinator, Orange Collective and Uncorrelated Ventures. Its software parses tables, charts and complex layouts into structured, traceable data that AI agents can consume, targeting high-stakes sectors such as banking, healthcare and law. The seven-person startup will use the funds to hire researchers and engineers, expand sales and build proprietary datasets.