Reducto raises $75M Series B to expand document AI platform

Reducto has closed a $75M Series B round led by Andreessen Horowitz (a16z), a move that will accelerate its push to turn messy documents into LLM-ready data for enterprise applications. The fresh capital brings the startup’s lifetime financing to $108M.

Since launching an early Parse API that fuses conventional OCR with vision-language model capabilities, Reducto has focused on extracting structured signals from complex files — from invoices and legal contracts to clinical notes. The company says it has now processed more than 1 billion pages for customers across startups and large enterprises, and that monthly throughput has climbed roughly 6x since its Series A.

Investors in the round include a mix of returning backers and new capital: alongside the lead, existing supporters such as Benchmark and First Round Capital participated, and Phosphor Capital also joined. The company also confirmed the addition of Jennifer Li to its board, signalling stronger ties to seasoned product and growth leadership in the AI ecosystem.

The financing will be deployed across model research and product development. Reducto plans updates to its inference stack and ingestion tooling with the specific aim of improving accuracy on difficult formats and expanding automation workflows. The startup frames its roadmap around enabling downstream applications — automated document workflows, agentic systems that can read-and-act, and analytics pipelines that feed LLMs with reliable context.

Reducto’s announcement comes as demand for solutions that tame unstructured enterprise data is rising. Companies attempting to operationalise LLMs are increasingly focused on the pre-LLM layer: reliable extraction, entity linking and provenance. Market demand for document-understanding solutions is accelerating, driven by regulatory reporting needs in finance, compliance in healthcare, and productivity gains across knowledge work.

To broaden access, Reducto is introducing a more flexible pricing model and a credit programme aimed at startups and academic researchers, intended to make the same infrastructure used by large customers available to smaller teams. The firm also reiterated recruiting plans across engineering, product and commercial functions as it scales.

Founders thanked customers, early supporters and their internal team for helping reach this inflection point, while stressing that the company’s mission remains focused on delivering practical, measurable impact from AI on real-world data. With this round, Reducto aims to cement its position as an ingestion layer for enterprise AI stacks and to push further into high-accuracy, production-grade document intelligence.