HyperAIHyperAI

Command Palette

Search for a command to run...

Felicis leads $7.5M seed for Unsiloed AI document tech

Unsiloed AI has raised $7.5 million in seed funding led by Felicis to help enterprises make complex documents usable by AI agents. Y Combinator, Orange Collective, and Uncorrelated Ventures participated, along with angel investors Aarthi Ramamurthy and Beth Turner. The San Francisco-headquartered startup is targeting a persistent enterprise AI problem: business documents often contain tables, charts, and complicated layouts that general AI systems misread, and errors can be costly for banks, hospitals, and law firms that rely on such records. Unsiloed’s software breaks down complex documents, extracts information from tables, charts, and other elements, and converts it into structured data for AI agents. It also shows where extracted information originated in the original document, allowing businesses to verify results rather than trust AI output blindly. Co-founder and CEO Aman Mishra said large language models are inherently probabilistic and can introduce deviation and hallucination with serious downstream impact. Mishra founded Unsiloed with Adnan Abbas, his former classmate at IIT Kharagpur; the two joined Y Combinator’s Fall 2025 batch after developing the technology. The startup says its advantage lies in processing complex documents through “dual-stream” vision models that handle text and image data separately, combined with “heat map-based chunking” that identifies the most important document sections. The approach directs more computing power toward figures, numbers, and charts, while a specialized decoder accounts for industry-specific terminology that can carry different meanings across business contexts. Unsiloed is already working with customers in finance, legal services, and healthcare. Mishra said one bank uses the technology to process documents, a law firm has deployed it on its own servers for contracts and old case files, and voice AI startup Bolna AI uses it to ingest knowledge bases for voice agents. Unsiloed enters a crowded market that includes document-processing startups Reducto and Unstructured, as well as offerings from Microsoft and Amazon, and general-purpose AI models. Mishra pointed to a test involving 542 financial documents from an enterprise customer in which Unsiloed achieved 99.78% accuracy across roughly 84,000 extracted data points, compared with major AI models, Microsoft Azure Document Intelligence, and the customer’s in-house tool. The results have not been independently verified. Unsiloed also claims it outperformed competing tools on Allen AI’s public olmOCR-Bench, scoring 88.0 versus 66.0 for Reducto’s agentic system and 39.9 for Unstructured across 1,403 PDF pages covering tables, mathematical expressions, scanned documents, and complex layouts. Pricing starts at one cent per page. Unsiloed currently has seven employees and plans to use the funding to hire researchers and engineers, expand sales, and develop proprietary datasets.

Related Links