Why We Built a Local-First Trading Platform

TigerTradeBot runs on your computer. Not “runs in your browser backed by our cloud.” Your strategies, your trade history, your backtest results and your journal live in TigerTradeBot’s databases on your own disk. This was the first architectural decision we made, and everything else followed from it.

Here’s why.

Your strategies are yours

A trading strategy is intellectual property. If you’ve spent months developing and validating one, it may be the most valuable thing on your computer. On a cloud platform, that strategy sits in someone else’s database, governed by someone else’s terms of service, visible to someone else’s employees and to whoever compromises their infrastructure.

On your machine, it stays on your machine. We don’t hold a server-side copy of your strategies or your trade history to leak, subpoena, or train a model on.

The app does talk to our license server. It checks your license once a day, looks for updates, and, only if you opt in, reports win/loss counts for the community strategies you run (counts only, no trade details). Your strategies and trade history aren’t part of those requests.

No vendor custody of your data

Cloud trading tools have a quiet failure mode: the company pivots, gets acquired, or changes its terms, and your trade history and research are on their servers, not yours.

Local-first inverts this. Your data sits in TigerTradeBot’s data folder on your own disk, and the app’s Backup & Restore page makes a copy you control. Keep that backup somewhere safe, the same way you’d treat anything else that holds account credentials.

Strategy work runs locally

Building strategies, running backtests and reviewing your journal happen on your computer, not in a shared backend, so there’s no queue behind someone else’s backtest. Some things genuinely need a connection: broker connections, live market data, any cloud AI you choose to use, and the license check. TigerTradeBot checks its license when it starts and then once a day, with a 6-hour grace period if our server can’t be reached.

Local AI, same principle

The AI copilot follows the same rule as everything else. Point TigerTradeBot at any OpenAI-compatible local server (LM Studio, Ollama, or similar) and the model runs on your hardware. Your positions, your P&L and your strategy logic don’t have to leave your machine to be analyzed. Cloud providers are supported, with your own API key, but they’re opt-in, not the default assumption.

The copilot explains and analyzes. It does not predict markets or give financial advice. No model does that reliably, local or cloud, and we won’t pretend otherwise.

The trade-offs, honestly

Local-first has costs. You install updates yourself (the app tells you when one is available and installs nothing until you click). You’re responsible for your own backups. And your computer has to be running for your strategies to run. We think traders, people who by definition care about custody, risk and control, are exactly the audience willing to make that trade.

The short version

Your strategies and trade history stay on your computer. You can back them up. Strategy work runs locally. You can run the AI locally. And nobody, including us, sits between you and your broker.

That’s the platform we wanted for ourselves, so we built it. You can try it on the free 12-month paper-trading license: one per email address, no card required. Trading involves substantial risk of loss and is not suitable for every investor — but your data, at least, shouldn’t be one of the risks.