Data partnerships
Data you can
build on.
We identify relevant data owners, agree licensing and permitted use, and prepare datasets around your research requirements.
- Targeted sourcing
- Clear licensing
- Careful preparation
Real-world knowledge. New possibilities.
Our identity graph spans 2.2B people and 380M businesses. We connect AI labs with licensed data and our expert network.
Anonymization requirements are reviewed for each dataset and licensed use. Preparation can remove identifying details while preserving useful context. Data that uses stable pseudonyms still requires privacy safeguards.
Tens of millions of dollars’ worth of data provided to leading AI labs and research companies.
Nearly a quarter billion dollars, within a year of founding.
01 / Beyond the open web
In the systems businesses run. In the decisions people make. In years of work that have never been available to AI. Nyne brings the right sources and the right labs together.
02 / For AI labs
Tell us what your models need to learn. We help you find relevant datasets and collaborate with people who know the field.
Discuss your research needsData partnerships
We identify relevant data owners, agree licensing and permitted use, and prepare datasets around your research requirements.
Expert collaboration
Work directly with specialists in the fields your models need to understand. Our expert network brings their knowledge and judgment into your lab’s research.
Every data partnership begins with ownership, permitted use, and agreed terms.
03 / For data partners
You’ve built something valuable. A data partnership should respect that. We work directly with owners to explore what their data could make possible, and agree how it can be used.
Explore a data partnership Looking for your privacy choices?We work with the people who can authorize a partnership and review the rights associated with the proposed data.
The data, permitted uses, and commercial terms belong in an agreed license. The conversation comes before the handoff.
Preparation and privacy review depend on the source and intended use. We consider identifying information alongside the relationships that make the data useful.
Start with a conversation. We’ll discuss fit, rights, and next steps before any data handoff.
04 / The Nyne identity graph
We use our identity graph of 2.2B people and 380M businesses to connect a lab’s requirements with relevant organizations and data owners. It helps us find who holds the right data and who can authorize a partnership.
Identity resolution, person and company enrichment, and people search. The human context for your applications and agents.
Explore the API documentationHow Nyne responds
A new set of capabilities. A new frontier to test.
Test the model against more demanding tasks.
Where does it fall short?
Financial reasoning04 / Nyne connects
Our identity graph and expert network turn a capability gap into targeted connections.
Human experts
Finance specialists create and review reasoning tasks.
Software specialists design and review tasks that require multiple steps.
Engineers contribute troubleshooting tasks and professional judgment.
Domain knowledge, applied.Business owners
Business owners license relevant financial records and accounting workflows.
Business owners license product workflows and interaction traces.
Business owners license maintenance histories and operational records.
Owner permission. Agreed use.Address the gap through training and evaluation.
Nyne research
Nyne is a research and data company. We investigate the gaps that benchmarks reveal, and how expert knowledge, better tasks, and real-world data can help address them.
Our research directions
Turning domain knowledge into demonstrations, reasoning tasks, and feedback that help models learn from expert practice.
Making expert judgment explicit through criteria that help assess reasoning, decisions, and task performance.
Exploring how real business workflows can become richer settings for training and evaluating AI agents.
Studying the actions, decisions, and context behind real work in software, from an individual step to a complete task.
The Nyne expert network
Work directly with domain specialists. Our expert network brings research teams together with people whose knowledge comes from years of study and practice.
Find expertise for your labNyne is a research and data company—the data factory for AI. We bring AI labs licensed data and human expertise, and investigate how these can help address model capability gaps. Our identity graph of 2.2B people and 380M businesses helps us find relevant data owners; we qualify sources, arrange licensing, and prepare datasets. Labs can also work directly with domain specialists through our expert network. Our people data APIs remain available for developers and businesses.
Our research directions include expert supervision, evaluation and rubrics, agent tasks and environments, and computer-use trajectories. We start with a capability gap and investigate the expert knowledge, tasks, and real-world data that could help address it. Research collaborations define the question, scope, and relevant data rights together.
The identity graph helps us find relevant businesses and the people connected to potential data sources. The expert network enables labs to collaborate directly with domain specialists, bringing human knowledge and professional judgment into their research. The 2.2B people and 380M businesses describe graph coverage, not expert-network membership or licensed partners. Data partnerships require ownership review and an agreed license.
Start with your research objective and the data or domain expertise you need. For datasets, we work with you to assess suitable sources, ownership, licensing, preparation, and delivery. For expert collaboration, we discuss the specialist knowledge and experience your work calls for and identify relevant people in our network.
Businesses and other data owners with useful records, domain knowledge, or operational history may be a fit. You need the authority and rights to license the proposed data. We start with a conversation about what exists and whether it matches a lab’s needs.
Ownership, rights, and permitted use are reviewed as part of the partnership. Before delivery, anonymization requirements and privacy handling are scoped to the dataset, licensed use, and agreed terms. Preparation can include removing identifying details or replacing identifiers while preserving relationships across records. Data that uses stable pseudonyms still requires privacy safeguards; pseudonymization alone does not make data anonymous.
Yes. Our API documentation covers people search, identity resolution, person and company enrichment, and integrations for AI agents. Use the API docs link in the navigation to explore endpoints and get started.
We’re hiring
Build the research and infrastructure that bring human expertise and business data to AI labs. We’re hiring people who want to help shape how AI learns from the real world.
Explore careers at NyneThe next connection starts here.
A connected world. A new possibility.