Retab

Retab is a developer-focused artificial intelligence (AI) company that provides a platform and software development kit (SDK) for processing and extracting structured data from documents. The company was founded in 2024 by Louis de Benoist (CEO), Sacha Ichbiah, and Victor Plaisance, engineers from Cambridge and École Polytechnique. The idea for Retab emerged from the founders' experience building internal automation tools for document-heavy workflows in the logistics sector.

The platform is designed to convert unstructured inputs like PDFs, handwritten scans, and emails into clean, structured data. It serves developers in industries such as logistics, finance, and healthcare. Retab operates as an intelligence layer on top of large language models (LLMs) from providers like OpenAI, Google, and Anthropic, making them reliable for high-stakes, real-world applications. The business model is centered on providing this developer platform and SDK, which manages the entire document extraction lifecycle, including dataset labeling, evaluation, automated prompt engineering, and model selection. Users define their required data schema, and the platform handles the rest, intelligently routing tasks to the best-performing model for the job.

In July 2025, Retab announced a $3.5 million pre-seed funding round backed by investors including VentureFriends, Kima Ventures, K5 Global, and notable individuals like former Google CEO Eric Schmidt. This funding is intended to expand the platform's capabilities to meet growing demand. The company has headquarters in both Dover, Delaware, and Paris, France.

Keywords: document data extraction, structured data, AI agent, developer platform, large language models, document processing, automation tools, PDF extraction, data schema, prompt engineering, model selection, logistics tech, fintech, health tech, unstructured data, AI infrastructure, API, SDK, document automation