Binance Quant Engine ⚙️

A strategy-agnostic Binance USDT-M futures backtest & execution engine. Backtest and live trading run through the same strategy code, so a look-ahead bug can't exist in one and not the other — and an MCP server lets any MCP-compatible agent run backtests without writing a line of Python.

Extracted from the infra layer of a real, currently-running Binance USDT-M futures bot — the alpha (strategy, parameters, live P&L) stays private; what's public is the part every retail algo trader rebuilds badly at least once: a backtester that can't cheat, and an execution layer that survives Binance's actual API quirks. The bundled demo strategy is textbook logic (Bollinger squeeze); it exists to exercise the engine end to end, not to make you money.

This repository supersedes quantbox-engine, which is no longer maintained. If you have that repo cloned, switch to this one.

Why not just write your own backtest loop?

Because the failure modes here are the ones that quietly wreck a real account, not the ones a unit test catches:

Quick start

pip install binance-quant-engine
# or: uv add binance-quant-engine

bqe-backtest --demo      # backtest the bundled demo strategy on synthetic data

⚠️ The numbers above are from a demo strategy on synthetic data. They prove the engine runs; they say nothing about profitability.

Backtest your own data

from binance_quant_engine.backtest.vectorized import run_backtest
from binance_quant_engine.data.klines import load_csv
from binance_quant_engine.strategy.demo_squeeze import SqueezeStrategy

high, low, close = load_csv("BTCUSDT_1h.csv")   # columns: high,low,close
result = run_backtest(SqueezeStrategy(), high, low, close, fee=0.0004, slippage=0.0002)
print(result.summary())
# {'n_trades': ..., 'total_return': ..., 'win_rate': ..., 'profit_factor': ..., 'max_drawdown': ...}

Bring your own strategy by implementing TradingStrategy — a typing.Protocol, no base class required. Wiring it into live execution: docs/USAGE.md.

MCP server — let an agent run your backtests

pip install "binance-quant-engine[mcp]"
binance-quant-engine-mcp   # stdio MCP server

Exposes run_demo_backtest, backtest_csv, and describe_strategy_protocol to any MCP client (Claude Code, Claude Desktop, etc.) — point an agent at a CSV of OHLC data and it can backtest a strategy idea in the same turn, no local Python environment required on the agent's side.

This is deliberately backtest-only. Nothing under execution/ (order placement, cancellation, bracket management) is reachable through the MCP server — there is no tool call that can touch a live Binance order. If you want an agent that also trades, that's a decision you wire yourself, explicitly, outside this server.

Design

Structure

Decide (strategy) / measure (backtest) / execute (live) are kept apart and meet only at TradingStrategy.

flowchart LR
    subgraph Data["binance_quant_engine/data/"]
        CSV[("OHLCV CSV\nload_csv()")]
        SYN[["synth_ohlcv()\nsynthetic data"]]
        CACHE["cache.py\nmemory + gzip cache"]
    end

    subgraph Decide["binance_quant_engine/strategy/ — decide"]
        PROTO{{"TradingStrategy\n(PEP 544 protocol)"}}
        SQZ["demo_squeeze.py\nSqueezeStrategy"]
        PROTO -.implements.-> SQZ
    end

    CSV --> BT
    SYN --> BT
    CACHE -.caches.-> CSV

    subgraph Measure["binance_quant_engine/backtest/ — measure"]
        BT["vectorized.run_backtest()\nhands close[:t+1] only\n(no look-ahead)"]
        RES["BacktestResult\ntrades · equity curve · MDD"]
        BT --> RES
    end

    BT <-->|"update_market_data / on_bar\nopen·update·close_position"| PROTO

    subgraph Execute["binance_quant_engine/execution/ — execute (.[live])"]
        HOST["host.py\nScalperProtocol (live bot host)"]
        BRACKET["brackets.py\nBracketMixin — SL/TP/trailing"]
        UTIL["utils.py\nround_to_tick · dead-status set"]
        ALGO["algo_api.py\nAlgoApiClient"]
        HOST --> BRACKET --> ALGO
        BRACKET -.-> UTIL
    end

    subgraph MCP["binance_quant_engine/mcp/ — optional (.[mcp])"]
        SRV["server.py\nrun_demo_backtest · backtest_csv"]
    end
    BT -.callable via.-> SRV

    PROTO ==same interface\n(backtest = live)==> HOST
    ALGO --> BINANCE[("Binance USDT-M\nFutures Algo Order API")]

• How to plug in a strategy and wire up live execution → docs/USAGE.md
• Design rationale for no-look-ahead, backtest/live parity, and the cost model → docs/ARCHITECTURE.md

License

MIT

Bugs and questions → Issues.

Younghwan Chae · GitHub @younghwan91 · LinkedIn — other open-source quant projects (Korean equities, US equities, crypto) are on the profile page.