Post 20 CompanyValency bond

From hallucination to citation: Valency MCP Bond comes to CBorg

Berkeley Lab researchers can now ground their AI work in the scientific literature, right where they already work.

Today I’m excited to share that Valency MCP Bond is now available in CBorg, the AI portal run by Berkeley Lab’s Science IT group. Berkeley Lab researchers were among our private-beta testers, so it means a lot to see Bond land in the portal they already work in.

What is CBorg?

CBorg gives Berkeley Lab researchers and staff secure access to commercial models from OpenAI, Google, Anthropic, and others under the Lab’s enterprise agreements, plus free open models hosted on the Lab’s computing infrastructure. Researchers use it for chat, summarization, coding, and analysis, or tap its API to build their own LLM-powered tools. CBorg brings these AI tools together in one place, making it easier for researchers to move from ideas to results.

Why do models need MCP Bond?

Modern AI models are powerful reasoners, but they were trained on yesterday’s science. Ask one for a citation and you might get a confident answer with no source, or a reference that doesn’t exist. For a scientist, “right-sounding” isn’t good enough.

Valency MCP Bond doesn’t try to out-reason your model. It makes sure the model is reasoning from the real, current literature. We continuously index papers and preprints from trusted sources, clean and cross-reference them, and make new work searchable within hours. Every result links back to a real source you can open and cite.

What Berkeley Lab researchers can do

With MCP Bond available directly in CBorg, Berkeley Lab researchers can put scientific literature to work in their AI workflows. They can:

  • Find relevant work faster with semantic search across papers, abstracts, titles, and venues.
  • Understand a field at a glance through publication trends, keyword trends, and emerging research domains.
  • Follow the people and the citations, including authors, coauthors, and the papers that cite a given work.
  • Take results with them by exporting to BibTeX, CSV, or JSON.

Instead of switching between a chat window and a search engine, they can ask a question and have their AI work from the real scientific record.

“Valency has transformed how I approach literature discovery, compressing what used to take days of manual searching into a few targeted queries taking minutes. Having cross-corpus semantic search and citation graph tools in a single interface has meaningfully accelerated the pace at which I can move from a research question to a well-grounded reading list to scientific discovery.”

— Dr. Peter Nugent, Senior Scientist and the Division Deputy for Science in the Applied Math and Computational Research Division at Lawrence Berkeley National Laboratory (LBNL)

Why Berkeley Lab, and why open standards

Valency’s roots run through Berkeley. Our CEO, Josh Bloom, is a UC Berkeley astronomy professor, and our advisor Fernando Pérez, co-founder of Project Jupyter, is a professor of statistics at Berkeley. We designed MCP Bond alongside researchers at national laboratories. The Lab is a world-class home for the interdisciplinary, data-heavy science that Valency was built to help accelerate.

Access is only half the story. How it gets delivered is just as important. MCP Bond is built on the Model Context Protocol (MCP), an open standard that lets one data layer serve many models without being rebuilt for each. That’s why we joined the Linux Foundation’s Agentic AI Foundation in August, and offering availability through a multi-model portal like CBorg reinforces why open protocols are essential for science.

Getting started

MCP Bond is available now to Berkeley Lab researchers through CBorg. Ready to try it? Visit CBorg and select Start Chat to access the AI portal. For information about using MCP servers, see the CBorg MCP documentation.

Once Bond is enabled, try asking your AI assistant:

“Use Valency to find recent research papers on AI applications in materials science, summarize the key findings, and provide citations to the original sources.”

Have feedback or questions? We’d love to hear from you at support@valency.io. Institutions interested in bringing Bond to their own researchers can contact sales@valency.io.

Thank you to the Berkeley Lab Science IT team for making this possible.


About Valency: Valency is the infrastructure that keeps AI working from the real record, not a plausible-sounding approximation of it. Founded by researchers and led by a practicing scientist, Valency provides people-centered, AI-accelerated research tools, including Valency MCP Bond, the trusted literature layer for AI assistants. Learn more at valency.io.

About Lawrence Berkeley National Laboratory (Berkeley Lab): Lawrence Berkeley National Laboratory (Berkeley Lab) addresses the world’s most urgent scientific challenges by advancing sustainable energy, protecting human health, creating new materials, and revealing the origin and fate of the universe. Founded in 1931, Berkeley Lab’s scientific expertise has been recognized with 17 Nobel Prizes. The University of California manages Lawrence Berkeley National Laboratory for the U.S. Department of Energy’s Office of Science.