Case Study: TrueNorth Trusts Cambrian Data for the World’s First Agentic Brokerage

This guest post is co-authored by Alex Lee and Willy Chuang, co-founders of TrueNorth.

At the intersection of AI, DeFi, and traditional finance, we now have the tools to allow anyone to build their own professional trading desk. This is what inspired TrueNorth, what I call the world’s first Agentic Brokerage. With TrueNorth, you can connect your Hyperliquid portfolio and get what only institutional quant funds used to have. It’s a desk of agents that surface your setups, question your thesis before you fire, watch your positions around the clock, and review every trade so the next one is even smarter. In other words, TrueNorth uses the world’s best data sources and a proprietary AI stack to compound your edge.

A real trading desk is more than an execution screen. It is a research floor, a risk function, and an intelligence layer that reads the market before any capital moves. TrueNorth gives you all of it as agents, and much of the data those agents read the market through is provided by Cambrian. In practice, you get visualized data, a written explanation, and a terminal to decide your next moves. This is what TrueNorth looks like:

TrueNorth in action on Hyperliquid: the agent returns a structured plan with momentum, trigger, and invalidation for the trader to act on.

TrueNorth’s core belief is that the problem was never simply access to signal. Most traders are not losing because they missed an opportunity. They are losing to themselves: overtrading, adding to the wrong side, repeating the same mistake at 3:47am. Crypto Twitter already throws twenty setups a day at everyone. At some point, more signal just becomes noise.

But for an agent to be your professional trading desk, it still needs to see the market clearly. To shape a setup around your edge instead of blasting the same call to everyone, and to watch your open positions for the one change that actually matters, it needs a clean, reliable, agent-native read on the market. That read is the data layer, and that is where Cambrian comes in.

Our agents increase their intelligence with a mix of data sources. One source of financial intelligence is Cambrian. Six of our production capabilities are powered by one provider, Cambrian:

Each of these capabilities represent one API call the agent makes when it needs to understand a token. It fans out one to three Cambrian requests in parallel and gets back a clean, structured read it can act on. Across the six capabilities we hit eight distinct Cambrian endpoints, and every response comes back in a stable, predictable shape.

That last part matters more for an agent than it ever did for a human. A person can squint at a messy chart and still get the gist (or imagine that it makes sense). When an agent has to choose between four overlapping providers for “holders,” reconcile mismatched shapes, and handle one of them being down, it spends its budget on plumbing instead of judgment. Cambrian collapses that. One provider allows us to tackle every question we would actually ask of a token.

Before a token can become a setup, the trading desk has to know it is real.A trader points the agent at $WIF, and Cambrian’s security endpoint returns a full contract audit: mint authority, freeze authority, balance mutability, metadata, transfer-fee and transfer-hook checks, each showing as revoked or none. Below the audit, holder concentration across the token’s 200,000-plus holders. This is the research a desk does before it commits a dollar, and every row is Cambrian on-chain data.

The data layer at work: a full contract-security audit and holder concentration, from Cambrian’s /tokens/{address}/security endpoint.

Then it reads the ownership.Ask how a token’s holders are structured and the agent reads Cambrian’s holder data back as tiers: Whale, Dolphin, and Shrimp, each with its wallet count, holdings, USD value, and 7-day trend. A concentration snapshot shows the top 10 wallets holding roughly 46% of supply and flags the largest whale adding 10.5M WIF in a single week. This is the market structure a desk weighs before it sizes a position, delivered in one pass.

Holder concentration and Whale / Dolphin / Shrimp tiers, read straight from Cambrian’s onchain data.

Returning to a point I made earlier: the bet behind TrueNorth is that the bottleneck was never the signal. It was the trader and the force of human error. The edge we are building is a desk that makes you self-aware: it surfaces setups shaped by how you actually trade, debates your thesis in the half-second before you click, watches your positions when you cannot, and after every close it files the trade and finds your real patterns. It’s a reflection you cannot lie to.

For any of that to work, the agent has to see the market the way a desk does. The moment it shapes a setup or weighs a position, it is reading structure: what a token is, who holds it, how the crowd is moving. That intelligence is sourced from Cambrian.

Self-custody runs through all of it. No deposits, no withdrawals, no markup on your fills. The agent observes, analyzes, and surfaces a judgment, and you are always the one who presses execute.

“At TrueNorth, we believe in sovereign capital. That means we don’t fully hand off funds to agents. We believe in human conviction, so the first step is rooted in human theses. Agents then amplify and enhance human intuition prior to executing trades.” — Willy Chuang, Co-Founder, TrueNorth

Money is fundamentally personal. The era of fully delegating capital to AI is not here, and it will not be soon, because the trust cost is too high. Capital sovereignty stays with the trader, and the agent only amplifies their judgment. For that judgment to be worth amplifying, the market read underneath it has to be clean. That is the standard Cambrian meets.

What made working with Cambrian unusually easy is that they envision the same future we do.

Their API was shaped for AI-native consumption from day one. Anyone who has done integrations feels the difference the moment they touch it. When we have had questions about response shapes or asked for new dimensions, the team has been fast and direct, the kind of context you only get from a partner who is values-aligned on the vision of the future of agents and finance.

“TrueNorth is building the first Agentic Brokerage, where every trader runs their own professional desk. Cambrian is the data layer that helps TrueNorth read the market through token security, holder structure, social momentum, and on-chain signals, all in one clean, agent-native API. Cambrian makes it easy for TrueNorth to give an agent the market truth it needs to shape setups around each trader and watch their positions in real time.” — Alex Lee, Co-Founder, TrueNorth

If you are building an AI-first product on crypto data, Cambrian is the API we would point you at.

We’ve launched TrueNorth Agents, empowering traders to create personalized AI agents that scan the market, set up trades, and watch your portfolio non-stop. We think it’s the best way to experience the power of agentic finance. As our friends at Cambrian are well aware, the best financial decisions are predicated on the best financial intelligence. When you amplify that decision-making with autonomous agents, that point rings true a hundredfold.

TrueNorth is the world’s first Agentic Brokerage, a personalized AI trade desk that gives every trader the analytics intelligence of an institutional fund. The platform combines structured reasoning, real-time data, and proprietary models to deliver financial insights that are accurate, context-aware, and execution-ready.

Cambrian is the financial intelligence layer for agents and institutions. Our API delivers real-time and historical blockchain data, covering yield, liquidity positions, risk, trading activity, and market sentiment, for agentic and institutional DeFi applications. Founded in 2024, Cambrian is backed by Polychain Capital, Franklin Templeton, a16z crypto, Flow Traders, Selini Capital, and others.