Retab makes any document AI-readable, emerges from stealth backed by $3.5M

Retab, the AI agent that builds document extraction pipelines, is announcing the launch of its innovative new platform backed by $3.5 million in pre-seed funding. Aiming to reimagine document processing and automation amid the era of large language models (LLMs), Retab allows developers to extract data from any document format, delivering a complete developer platform and SDK.

Underscoring the launch of Retab is the founders’ focus on the “broken state of document AI,” defined by inefficient internal automation tools for document-heavy workflows. Through their work, they found that the answer to revitalizing document processing is the orchestration layer, not simply improving the output itself.

“People keep building demos that look like magic, but break the moment you put them into production,” said Louis de Benoist, co-founder and CEO of Retab. “We lived that pain ourselves. Wiring up fragile pipelines just to extract a few fields from a PDF. We built Retab because it’s the developer-first platform we always wished we had.”

Retab is an intelligence layer that sits between popular AI models—including OpenAI, Google, and Anthropic—and enterprises’ unstructured data, making it usable for critical workflows, according to the company. The developer defines the data they need, and Retab manages the entire lifecycle—from dataset labeling and evaluations to automated prompt engineering and model selection.

“Retab is the OS for reliably extracting structured data,” said de Benoist. “It wraps the best models in a layer of logic that actually makes them usable with error handling and structured outputs. That’s what devs need if they want to build production apps, not just prototypes.”

Some of Retab’s capabilities include:

• Self-optimizing schemas involving an AI agent that automatically tests and refines instructions based on a user’s documents to increase accuracy before the system begins operation
• Intelligent model routing within a model-agnostic platform, automatically benchmarking and routing tasks to the most appropriate model for the job, depending on the priority—whether cost, speed, or accuracy
• Guided reasoning and k-LLM consensus that forces models to “think” step-by-step, implementing a consensus mechanism to quantify uncertainty across multiple models

With its latest infusion of capital—in a funding round backed by VentureFriends, Kima Ventures, and K5 Global—Retab drives platform development and community growth, scaling its infrastructure to meet rising demands, according to the company.

“The AI-fication of the economy depends on the capability to convert operations based on millions of documents into verified, structured data that autonomous systems can utilize. On a large scale, this process hinges on quality control, cost efficiency, and rapid implementation,” explained Florian Douetteau, co-founder and CEO of Dataiku and investor in Retab. “The team at Retab understands this thoroughly and is uniquely positioned to solve it for the thousands of AI first companies that are emerging.”

To learn more about Retab, please visit https://www.retab.com/.