The policeman was unimpressed by the sale deed. Sanjay Mandava had returned to India to deal with an illegal structure on his land and assumed the signed record of purchase would settle the matter. It did not. A sale deed recorded a transaction between two parties, he was told; it did not, by itself, prove title. To establish ownership, Mandava needed a small library of other documents from offices he had never learned to navigate.
It was a maddening answer, but also a precise product brief. Somewhere across registration departments, revenue records, municipal files, surveys and old transactions, the evidence existed. The owner’s problem was gathering it, reading it and knowing what mattered. Mandava began pulling documents from the states where his family held property. The first version of Landeed was built to answer his own question: what does the record actually say?
The company that grew from that search now describes itself as property intelligence infrastructure. It retrieves records for ordinary owners and helps enterprises conduct title searches, collateral checks and due diligence. Its newer Terra product lets a user ask questions in plain language and receive answers grounded in the underlying documents. The ambition is grand. The origin is beautifully inconvenient: a founder standing in front of the police, holding a paper that was both important and insufficient.
The dinner table was the first data room
Mandava moved from India to Chicago at nine. His family’s land stayed behind, and stories about it travelled back across the ocean. At dinner, he heard of disputed boundaries, attempted encroachments and the exhausting work of managing property from another country. The parcels represented inheritance, memory and plans that never quite made it off the drawing board. They also revealed a strange imbalance: an asset could be valuable on paper and still feel dangerously difficult to defend.
For a long while, that problem remained background noise. Mandava entered finance and worked as an investment analyst at Knight Vinke, an activist investment firm. Public markets offered price feeds, filings and the seductive neatness of a screen. Property offered a fence, a document in an unfamiliar format and several opinions about what might be true. He has recalled earning more money than he had time to spend, but a call from home made the old land problem impossible to keep at a polite distance.
His response was characteristically physical. He bought additional parcels and built mechanized warehouses, partly because putting something useful on land is one practical way to protect it. The warehouses pulled him into the movement of goods. That, in turn, led to GoLorry, an online trucking marketplace he built with ZJ Lin and Jonathan Richards. The trio joined Y Combinator’s Winter 2016 batch and tried to bring instant pricing and digital coordination to India’s fragmented freight market.
Mandava moved back to India full-time in 2013. GoLorry taught the team what tidy software diagrams tend to conceal: the physical economy is made of regional practices, phone calls, paperwork, people and exceptions. A truck does not care that the dashboard looks elegant. A land record is equally indifferent. It either matches the parcel, owner and history, or it does not.
A second company with the same co-founders
Landeed reunited the GoLorry trio around another stubborn corner of India’s physical economy. Mandava became CEO, Lin led product and Richards joined as a technical co-founder. Their second Y Combinator batch, Summer 2022, came with an advantage no accelerator could manufacture: years of knowing how each other worked when the plan met reality.
The early consumer proposition was disarmingly direct. Enter details such as an area or survey number and retrieve official property documents without the familiar tour of agents and offices. An encumbrance certificate could reveal transactions and claims across decades. Other records could identify ownership, land classification, taxes or changes to a parcel. Landeed stitched together systems that had been designed state by state, department by department and, often, era by era.
The useful answer rarely lives in one office. Property intelligence comes from connecting records that were created for different purposes.
The company announced an $8.3 million seed round in January 2023, co-led by Draper Associates, Y Combinator and Bayhouse Capital. It expanded from document access into legal support, enterprise due diligence and AI-assisted analysis. By late 2025, Mandava said Landeed ran millions of searches each month. Its pitch had evolved from a faster document window to something closer to an operating system for ownership.
The distinction matters. Digitizing a registry can mean little more than scanning its errors. A searchable mistake is still a mistake. Mandava argues that confidence requires connected systems, common standards and the ability to trace an answer back to the record. His preferred metaphor is a “blue tick for property,” verification made visible before money changes hands.
Build the records before the magic
Terra, launched through Y Combinator in 2026, is the clearest expression of Landeed’s next act. The company says its engine works across more than 500 million title records, more than 20 states and 22 languages, including handwritten and scanned material. A user can ask who owned a property, whether an encumbrance appears, how two options compare or what recent rules might affect a parcel.
One example has the useful oddness of real life. A user described a Chennai property by saying it was next to a biryani shop. Terra identified the place, pulled the records and found a recent municipal announcement that allowed additional square footage on the plot. The user had not known to ask. Search found the requested property; connected records found the more valuable question.
This is where Landeed’s years of document work become strategically important. The fashionable route to an AI product begins with a model and searches for data. Mandava’s route began with the data nobody enjoyed collecting. The model arrived after the team had lived inside the workflow. Its claim to usefulness rests less on conversational polish than on whether the conversation can point back to something official.
That sequence is the most portable idea in Mandava’s story. Founders often automate the visible complaint. Landeed insists on learning the hidden process first. The approach is slower at the beginning and harder to imitate later. It also makes a customer promise refreshingly concrete: the app will not claim to do what the field team cannot do.
Hyderabad, high bars and unfinished work
Mandava chose Hyderabad after trying several Indian cities. He liked the infrastructure, airport access, lower operating costs and depth of technical talent created by large global technology offices. The choice also kept Landeed close to one of its strongest markets. He has said Hyderabad accounts for roughly 80 percent of the company’s Telangana business, a fitting base for a product obsessed with the texture of local systems.
His public writing is direct, impatient and preoccupied with workarounds. He notices the person employed to press elevator buttons, the person hired to explain a form and the account manager required to make software feel reliable. His complaint is not with the worker. It is with the unfinished system that makes permanent human patching necessary. When a process never becomes dependable, relationships replace standards and “who you know” becomes infrastructure.
That impatience carries into hiring. Mandava talks about clarity of thought, functional skill and a hunger to learn. Landeed’s public recruiting language prizes builders and gives unusually broad invitations to people whose ability may not fit a conventional title. A discussion about founder stubbornness became the theme of an entire interview. The useful version of that trait appears in the company’s long attention span: two accelerator batches, years of field work and an insistence that property records deserve the same product seriousness as payments.
Watch: Property rights, revisited
In a 17-minute TEDx talk, Mandava examines how constitutional change shaped the status of property rights in India.
Play the TEDx talk ↗The search box is only the surface
Mandava’s aspiration extends beyond faster paperwork. He wants a property system in which an owner, buyer, bank or developer can start with evidence rather than connections. In his telling, clearer ownership is both economic and civic infrastructure. Families protect their largest asset. Lenders understand collateral. Developers move with fewer surprises. States collect cleaner revenue. The humble title search becomes a trust machine.
There is also a private loop still waiting to close. The land disputes that filled his family’s Chicago dinners interrupted plans they once had for those parcels. Mandava has said he wants to help complete that deferred family project someday. The aim gives Landeed’s scale a human unit: millions of records, yes, but ultimately one family able to act on its own land without fear or folklore.
In the meantime, the work is gloriously specific. A faded scan must be read. A survey number must match. A municipal notice must reach the person who did not know it existed. The original sale deed remains important, only no longer lonely. Around it, Landeed is assembling the rest of the story.
Follow the trail
Explore Mandava’s work, public writing and longer conversations about property, technology and building in India.