companies

Reducto

Document intelligence API that turns messy PDFs, scans, and spreadsheets into structured, LLM-ready data using vision models plus OCR.

In inglese

Reducto was founded in 2023 by MIT classmates Adit Abraham and Raunak Chowdhuri, who built the first version of the product over a weekend before enterprise customers started asking for it. The company is headquartered in San Francisco and has raised over $100 million across seed, Series A, and Series B rounds from Benchmark, a16z, First Round Capital, and Y Combinator. Reducto's platform combines computer vision, layout-aware models, and vision-language models to parse complex documents that break traditional OCR: dense financial tables, nested forms, scanned contracts, and mixed-language filings. Its Agentic OCR feature runs a VLM agent over baseline OCR output to catch and correct errors before the data reaches a customer's pipeline, and the API exposes separate parse, extract, split, and edit endpoints so teams can pull structured JSON without building their own post-processing. The company counts Harvey, Scale AI, and Vanta among its customers and says it has processed more than a billion pages, mostly for finance, legal, and healthcare teams building retrieval and extraction systems on top of large language models.

Founded

2023

Headquarters

San Francisco, United States

Sector

data