About

You Can Just
Do Things

We spent twenty years in the wilderness so you wouldn't have to.

What started as a frustration with blind-spot trading became an obsession with geometric truth. Traditional market analysis felt like navigating by looking in the rearview mirror—always late, always guessing. We wanted sensors that could see the terrain, not just the tracks left behind.

So we built them.

> The Problem Was Always: The Math Didn't Exist Yet

Here's what they don't tell you about innovation: sometimes you have to invent the tools before you can build the thing.

This didn't start with "let's apply topology to finance." It started with a paper on holomorphic cohomology and a gnawing feeling that the geometry of complex systems could reveal something markets were hiding.

But the framework wasn't there.

Topological Data Analysis as a formal field? That came later. What we had were concepts—persistent homology buried in pure mathematics, particle filters used in robotics, quaternion algebra from aerospace navigation. We had pieces of a puzzle scattered across disciplines that had never spoken to each other.

So we built the bridges.

Years of:

  • >Restless nights working through problems that didn't have published solutions
  • >Trial and error with particle filters to approximate topological features before the formal tools existed
  • >Reading papers from quantum physics, differential geometry, robotics, control theory
  • >Building, breaking, rebuilding mathematical frameworks that could handle the chaos of financial data

By the time "Topological Data Analysis" became a recognized field, we'd already been running proto-TDA methods on live market data for years.

We didn't wait for permission from academia. We just started building.

> Standing on Giants' Shoulders—Then Climbing Higher

The giants gave us the raw materials:

  • Quaternion algebra (Hamilton, 1843) for encoding rotational states
  • Persistent homology (Edelsbrunner, 2000s) for tracking topological features
  • Forward kinematics (Denavit-Hartenberg, 1955) for robotic motion planning
  • Particle filters (Gordon et al., 1993) for real-time state estimation

But none of them were designed for finance.

Our innovation wasn't "what if we used topology?"

Our innovation was:

  1. 1. Figuring out how to encode OHLC data as quaternion states when the literature said it couldn't be done efficiently
  2. 2. Adapting particle filters to approximate topological features in real-time before computational TDA became feasible
  3. 3. Mapping forward kinematics to momentum tracking through multi-dimensional data space
  4. 4. Synthesizing all of it into a coherent framework that doesn't require a PhD to use

This is physics-based finance.

Not "inspired by physics"—literally using the same mathematical frameworks that describe particle motion, electromagnetic fields, and quantum state evolution. Applied to market microstructure.

The giants solved universal problems. We solved the translation problem.

> What Was Impossible Is Now Infrastructure

The breakthrough wasn't just mathematical—it was making it scalable.

Twenty years ago, running real-time topological analysis on market data would have required a research lab's budget, custom hardware, a team of specialists, and institutional clients to justify the cost.

That barrier kept this locked in the realm of theory.

But infrastructure evolved:

  • Edge computing made distributed computation cheap
  • LLMs made complex outputs interpretable
  • On-chain execution made custody separation trivial
  • Cloud infrastructure made reliability accessible

Suddenly, what required a fortress could be delivered as an API.

  • Quaternion state encoding that was theoretically elegant but computationally brutal? Now <50ms per symbol via streaming.
  • Topological precursor detection that needed approximations and prayers? Now deterministic and reliable.
  • Kinematic momentum tracking across thousands of assets simultaneously? Now a standard API endpoint.

The math is rigorous. The infrastructure is battle-tested. The API is simple.

> What We Actually Believe

1. The Best Tools Should Feel Obvious in Hindsight

Complexity belongs in the math. The interface should be clean enough that your AI agent can figure it out without reading a 200-page manual. With LLMs, you don't need to teach users topology vocabulary—Claude can translate geometric features into plain English.

2. Co-Pilot > Autopilot

We're not here to trade for you. We're here to give you sensors that see what humans can't. You bring the strategy, the intuition, the risk tolerance. We bring the topology. You stay in control.

3. Physics-Based Doesn't Mean Theoretical

Every framework we use has been proven in the real world—just not in finance. Quaternions guide spacecraft. Particle filters track fighter jets. Forward kinematics moves robotic arms with millimeter precision. We translated field-tested math to market-tested systems.

4. Open Beats Proprietary (Eventually)

Lock it up long enough to prove it works. Then open the gates. The real value isn't in hiding the method—it's in executing it better than anyone else. If you can rebuild this, go for it. We're betting our implementation is good enough that you won't bother.

5. Custody Is Sacred

We never hold your keys. We never see your positions. We stream topology; you execute trades. Separation of concerns isn't just good architecture—it's the only ethical way to build in DeFi.

> What This Enables

For Quants

Stop reinventing topological analysis from scratch. We spent twenty years on this. Query the framework and get back to building strategies.

For AI Agents

Your LLM can now "see" market structure the way a physicist sees phase space. Ask geometric questions, get structured answers, execute with precision.

For Traders

You don't need to understand Hermitian cohomology. You need to know when volatility is about to spike. We translate the math into signals you can act on.

For Builders

The first physics-based finance framework available as infrastructure. Build the next generation of on-chain strategies using geometric intelligence that was previously locked in research labs.

> The Journey Was Long. The API Is Fast.

Twenty years of reading papers, working through unsolved problems, building proto-frameworks, running approximations, and watching infrastructure evolve until the pieces finally aligned.

And now it's just an endpoint.

const topology = await fetch('https://api.kinematic.finance/v1/tda/precursor', {
  method: 'POST',
  body: JSON.stringify({ symbol: 'BTC', timeframe: '1h' })
});

That single call represents two decades of work. You get the results in milliseconds.

That's the point.

You shouldn't have to spend twenty years figuring this out. You should be able to just use it and build something incredible on top.

Try It

Connect your wallet. Get an API key. Query the topology.

See what happens when you can finally see the market's shape, not just its shadow.

Twenty years in the making. Milliseconds to execute.

You can just do things now.