Reducto, a Y Combinator backed startup that helps parse documents, has raised $24.5 million in a Series A funding on April 27. The startup makes use of a mix of traditional computer vision and new Vision Language Models (VLMs) — to help companies turn their most complex documents into precise, LLM-ready inputs.
Reducto has become the go-to solutions for AI teams in thousands of companies, including Scale AI and FAANG enterprises. The company cofounders and MIT alumni, Adit Abraham and Raunak Chowdhuri said they are focusing on making Reducto the definitive platform for leveraging unstructured data end-to-end.
Founded in 2023, Reducto has extended support to power comprehensive workflows — including document splitting, intelligent classification, precise structured extraction and more, with its upcoming platform integrating these capabilities to help enterprises build accurate pipelines with their unstructured data.
READ: AI chip startup FuriosaAI turns down $800 million acquisition offer from Meta (March 24, 2025)
This funding round was led by Benchmark, with contributions from existing investors First Round Capital, BoxGroup, and Y Combinator. This funding round brings Reducto’s total funding to $33 million.
Reducto has also launched two key improvements — a new agentic OCR framework, which automatically reviews Reducto’s outputs, catching mistakes and making corrections through a multi-pass VLM framework, similar to having a human in the loop, and smart cost savings for simpler pages.
Reducto has partnered with some of the largest enterprise companies across industries —finance, healthcare, tech, and legal — parsing over 250 million pages of documents to help solve critical bottlenecks for their AI teams.
From parsing millions of real-world documents to helping the best AI teams scale, Reducto is becoming foundational infrastructure for the next wave of AI, the company claimed on LinkedIn. Reducto’s customer base includes Airtable, Scale AI, Vanta, and Fortune 25 companies, parsing their documents with confidence and security.
Abraham, who’s cofounder and CEO has previously worked with Google, BlinkAI, and MIT Media Lab, while fellow founder Chowdhari was previously a machine learning researcher at MIT Driverless, and MIT Research Laboratory.