China's 4.8 million factories,
as one callable data layer.
Registry tools tell you who incorporated. B2B directories tell you who's running ads. Tianxia Gongchang tells you what a plant produces today, how much it can make, who signs off, and how to reach them.
PRODUCTS
What we build
Three lines: a platform you can use, an API you can call, and AI capabilities built into your systems.
In a trillion-dollar manufacturing market, search remains an unsolved problem — not because search boxes are bad, but because this industry's customers were never indexed by Google.
In-Equal engineering · internal memo
NEWS
Recent updates
4.8 Million Factories Were Already in the Database — The Hard Part Was Teaching the AI to Find the Right Twenty
Over the past six months we turned factory sourcing into a conversation. Reviewing all of it, not one difficulty came from the data — what we needed was always in the database. The hard part was teaching the AI to turn what a customer says into the right question to ask of it.
— In-Equal Editorial
The Leaderboards Turn Over Every Six Months — Why We Are More Committed Than Ever to Vertical AI
In the first half of 2026, frontier models iterated at the pace of a new generation every two months, and on coding benchmarks the gap between open-source and closed-source models effectively closed to zero. The stronger, cheaper, and more homogeneous models become, the more the value of an AI application is determined by things outside the model — data, workflows, and understanding of the industry.
— In-Equal Editorial
Your Next Customer May Have Asked AI First: Manufacturers and the Problem of Visibility to AI
A Forrester survey shows 94% of B2B buyers use AI in their purchasing process, and search engine clicks are flowing into AI answer boxes. When customers start asking AI "which manufacturers in this category are reliable," manufacturers need to answer a new question: does AI know you?
— In-Equal Editorial