Download the Mac app, connect your AI agent, and run your first researched, out-of-sample-validated strategy — in minutes, no code required.
Download RLX Backtester for macOS, drag it to Applications, and launch it. On first run, paste your license key — the app starts a local engine and is ready to drive.
Download RLXBT for Mac ↓Requirements
macOS on Apple Silicon (M-series). The app bundles its own Rust engine — nothing else to install. Your license key activates it; it's verified online with 24h caching for offline use.
The running app exposes an MCP endpoint. Point Claude Desktop, Cursor, or any MCP client at it and the agent gets 30 tools — load data, backtest, walk-forward, Monte-Carlo, optimize, train RL, predict.
http://127.0.0.1:8142/api/mcp/sse Open your Claude Desktop config file (~/Library/Application Support/Claude/claude_desktop_config.json) and append:
{
"mcpServers": {
"rlxbt": {
"url": "http://127.0.0.1:8142/api/mcp/sse"
}
}
}Go to Cursor Settings → Features → MCP, click + Add New MCP Server, and input:
Then just talk: "Backtest an RSI idea on BTC and prove it out-of-sample, then train an RL agent." Everything renders live in the app while the agent works.
That's the whole setup. Now ask your agent in plain language — it picks the right tools, runs the work, and everything renders live in the app.
"Load my BTC data, backtest an RSI mean-reversion idea, prove it out-of-sample with walk-forward, then train an RL agent and tell me the current signal."
Curious what else it can do? See the full list of agent capabilities.