Skip to main content
R

Reducto

Reducto provides an agentic document platform using OCR and vision language models to convert messy real-world documents into structured data for enterprise AI teams.

Reducto builds an agentic document platform that turns messy, real-world documents - scanned forms, complex financial statements and the like - into structured, actionable data for enterprise AI teams. Its document processing toolkit works through documents the way a human would, using a multi-pass system that combines custom in-house models with leading frontier models, OCR and vision language models. The approach is designed to deliver accuracy and reliability on the kinds of documents that defeat conventional extraction pipelines.

The company's technology spans OCR, vision language models, document understanding, agentic AI and structured data extraction, with in-house model development alongside frontier models. Customers and partners number in the hundreds and include firms in legal (Harvey), compliance (Vanta), AI data labeling (Scale AI), financial services - among them top trading firms - and big tech. Revenue has grown eightfold year over year, and the company has raised over $100m from investors including Andreessen Horowitz, Benchmark and First Round Capital.

Reducto describes its mission as unlocking messy real-world data to power AI. Work centres on enterprise performance and scale, with the engineering challenges spanning model development, multi-pass document processing and the infrastructure required to serve demanding enterprise workloads.

Open jobs

No open jobs right now.