AI Workflows Get New Open Source Tools to Advance Document Intelligence, Data Quality, and Decentralized AI with IBM’s Contribution of 3 projects to Linux Foundation AI and Data

New projects strengthen the open source AI and data ecosystem and expand the Foundation’s technical portfolio 

SAN FRANCISCO, Calif., April 29, 2025 – LF AI & Data Foundation, an umbrella foundation of the Linux Foundation supporting open source innovation in artificial intelligence and data, today announced the induction of three new open source projects contributed by IBM: Docling, Data Prep Kit, and BeeAI. All three projects have officially been inducted by the LF AI & Data Technical Advisory Committee.

These contributions significantly enhance LF AI & Data’s technical landscape in three rapidly growing domains—semantic document understanding, enterprise-grade data preparation, and privacy-preserving federated learning—reinforcing the foundation’s mission to build a sustainable and open AI ecosystem.

The New Projects:

• BeeAI is the first open-source agent-to-agent platform for developers to build, discover, run, and compose agents and create multi-agent workflows. Powered by the open Agent Communication Protocol (ACP), BeeAI makes it easy to discover and connect AI agents from any framework or tech stack.
• Docling is an open-source, state-of-the-art ecosystem of tools (python packages) to do document conversion, generation and manipulation. It enables users to easily build pipelines to extract structured information from complex documents. With over 27K stars on github, Docling is already well on its way to becoming the de facto standard.
• Data Prep Kit is a modular suite of tools designed to clean, transform, and trace unstructured data for LLMs with a focus on quality, transparency, and scalability. It supports both batch and streaming data scenarios and integrates easily with modern AI workflows.

“We are excited to welcome Docling, Data Prep Kit, and BeeAI into the LF AI & Data family,” said Todd Moore, SVP, Community Operations at the Linux Foundation and interim Executive Director, LF AI & Data. “These contributions from IBM reflect a strong commitment to open collaboration and responsible AI. I love BeeAI’s commitment to both Javascript and Python for aggregated learning.”

“Docling, Data Prep Kit, and BeeAI were born from a need to fill critical gaps in AI development tooling and accelerate innovation in the Generative AI space. We’re proud to see them as a catalyst enabling the broader open-source community to build AI applications and agentic workflows,” said Brad Topol, Distinguished Engineer and Director of Open Source, IBM. "We’re excited to collaborate with the open-source community to evolve these technologies and solve real-world challenges together."

Governance & Community Collaboration

The projects will benefit from the governance, technical support, and ecosystem engagement that LF AI & Data provides to its hosted projects.  All three projects have officially been inducted by the LF AI & Data Technical Advisory Committee (TAC) and will establish neutral, community-driven technical steering committees.

The projects are now publicly available for exploration and contribution. Developers, data scientists, and researchers are encouraged to get involved and shape the future of these impactful technologies.

For more information and to get involved, visit: https://lfaidata.foundation