Manufacturing
Reducto, an AI-driven startup that turns unstructured data from PDFs and spreadsheets into formats that large language models (LLMs) can efficiently process, has raised $8.4 million in a seed funding round. The round was led by First Round Capital, with participation from prominent investors including Y Combinator, BoxGroup, SVAngel, and Liquid2. Notable angel investors such as Dropbox founder Arash Ferdowsi, Airtable founder Andrew Ofstad, Zeus founder Kulveer Taggar, WayUp founder JJ Fliegelman, WePay founder Richard Aberman, and PlanGrid founders Ralph Goottee and Tracy Young also joined the round.
Founded in 2023 and part of Y Combinator’s Winter 2024 batch, Reducto’s innovative solution addresses a critical pain point for AI teams across industries: the difficulty LLMs face when reading and extracting information from complex document formats. PDFs, in particular, often contain a mix of text, charts, tables, and images that are difficult for AI models to interpret accurately, leading to potential hallucinations and errors in output. With its advanced vision models, Reducto aims to bridge this gap by making these complex data sources easy for LLMs to read and process—similar to how a human would approach the same task.
“We just finished building what we believe is state of the art for spreadsheet parsing,” said Adit Abraham, CEO of Reducto, in an interview with Business Insider. “Our goal is to be the layer that connects human data with LLMs, enabling them to work with real-world documents at scale.” According to Abraham, Reducto’s models are capable of recognizing intricate visual cues in documents, such as spaces between paragraphs that signal topic changes or tabs that indicate nested hierarchies in lists.
