TradeAssembly

BUILD YOUR OWN AI TRADING FACTORY

Create a trading system
that can explain itself.

TradeAssembly is the operating layer for agent-driven trading systems: durable tools around your agent and broker for the work that needs to be explicit, constrained, and inspectable.

Your agent thinks. Alpaca provides brokerage and market access. You set the strategy, risk scale, account, and activation decisions.

WHAT YOU’RE BUILDING

An operating system around the trade—not another chat window.

TRADEASSEMBLY CORE

Local, open-source foundation.

Core gives your agent durable tools for strategy versions, backtests, deterministic risk checks, journals, receipts, reconciliation, recovery, and pause controls.

TRADEASSEMBLY RELAY

Optional paid companion.

Relay watches your Core node, retains evidence off-device, and supports secure remote inspection and supported controls.

WHY THIS LAYER EXISTS

A broker API can place a trade.
It cannot explain the trade.

For a simple experiment, an agent with the Alpaca CLI may be enough. But a CLI alone does not keep a durable record of which strategy version ran, which backtest supported it, which risk rules applied, whether an outcome was uncertain, or how to recover cleanly after a restart.

Core does not replace your agent. It gives the agent stronger tools and hard boundaries around the decisions that deserve more than prompts, chat history, and API calls.

WHAT CORE DOES

Give your agent a system of record.

Core makes the machinery around a strategy visible before, during, and after a run.

01

Version the intent

Define strategies and deployment configuration. Keep the exact version behind every run.

02

Keep the proof

Retain backtest inputs and result IDs so the case for a system is inspectable later.

03

Check the boundaries

Apply deterministic risk and authorization checks before an action is allowed.

04

Keep the receipts

Maintain a durable journal and exportable evidence trail, including uncertain broker outcomes.

05

Reconcile before retrying

Treat a timeout as a question to resolve—not permission to repeat a side effect.

06

Pause and recover

Record enough state to pause or deactivate a deployment and recover after a restart.

HOW THE FACTORY WORKS

Keep the roles clear.

  1. 01

    Your agent reasons.

    Use your own agent environment to turn an idea into strategy logic and deployment configuration.

  2. 02

    Core makes it reviewable.

    Version the strategy, retain backtest evidence, and apply deterministic checks before an authorized action.

  3. 03

    Alpaca provides market access.

    Your broker account and permissions remain yours. You decide whether and when a deployment is activated.

  4. 04

    Relay extends the view.

    Add off-device monitoring, retained evidence, backup, and supported remote inspection when you need it.

WHO IT’S FOR

For builders whose questions start after the trade.

TradeAssembly is for people building their own agent-driven trading systems who want more rigor than prompts, chat history, and broker calls alone can provide.

Request Core + Relay access

You should be able to answer:

  • Which exact strategy version produced this trade?
  • What research or backtest supported it?
  • Which risk policy allowed or rejected it?
  • Did the broker actually accept it?
  • What happened while my agent or computer was unavailable?
  • Can I inspect, export, or reproduce the evidence later?

WHAT RELAY ADDS

Watch it. Remember it.
Reach it.

Relay is the paid companion to your local Core installation. It gives you a way to see more than whatever happens to be on the machine in front of you.

Watch. Receive off-device alerts when a Core node is offline or falling behind.

Remember. Keep searchable journals, strategy versions, backtest artifacts, and configuration off-device.

Reach. Securely inspect status, history, and supported remote controls such as pause or deactivation.

Relay does not host your agent or keep your strategy running when your computer is off.

HOW TO START

Bring the ambition.
Keep the authority.

Start with a Mac running Core, an agent environment such as Codex, Claude Code, or OpenClaw, an Alpaca account with permission to connect, and your own strategy and risk rules.

Core makes the workflow durable, constrained, and inspectable. Add Relay when you need off-device evidence and access.

Request Core + Relay access