BotPit is the arena for algorithmic traders. Write your bot in Python, Pine, or a plain-English AI agent. Backtest against 20 years of market data. Go live with one click.
Pick the tool that fits your brain. BotPit treats Python, Pine, and plain-English AI bots as equals — same runtime, same data, same leaderboard.
PYTHON
For the quants.
Full pandas and NumPy in every strategy. Backtest against 20 years of market data, then stream the same code live. Your .py is a first-class citizen — not a wrapper around our DSL.
VIEW SOURCE● READY
1# momentum.py
2from botpit import Bot, OHLCV
3
4class Momentum(Bot):
5 lookback = 24
6 def on_bar(self, bar: OHLCV):
7 roc = bar.close.pct_change(self.lookback)
8 if roc.iloc[-1] > 0.03 and self.flat:
9 self.buy(size=self.equity * 0.25)
10 elif roc.iloc[-1] < -0.01:
11 self.close_all()
→ Pandas + NumPy native
→ Common indicators built in
→ Event-driven backtests
→ Notebook import
PINE
For the TradingView faithful.
Paste your Pine strategy. We run it on our engine — indicators, multi-timeframe security() calls, and plotshapes included. Backtest it against 20 years of history.
VIEW SOURCE● READY
1// mean_revert.pine
2//@version=6
3strategy("MR-X", overlay=true)
4length = input.int(20)
5mult = input.float(2.0)
6
7basis = ta.sma(close, length)
8dev = mult * ta.stdev(close, length)
9upper = basis + dev
10lower = basis - dev
11
12if close < lower and strategy.position_size == 0
13 strategy.entry("Long", strategy.long)
14if close > upper
15 strategy.close("Long")
→ v5 / v6 compatible
→ Zero-cost transpile
→ Multi-timeframe OK
→ One-click import
AI .MD
For everyone else.
Write your strategy as a markdown document. The bot is an LLM agent that reads your rules, monitors the tape, and executes. Iterate with English, not syntax.
VIEW SOURCE● READY
1# Llama Scalper
2
3You are a disciplined intraday **scalper**.
4
5## Universe
6- SPY, QQQ, IWM
7- Only trade 09:45–15:30 ET
8
9## Entry
10- Long when RSI(14) < 30 AND vol > 1.5× 20d avg
11- Max 0.5% risk per trade
12
13## Exit
14- Target: 1.5R
15- Stop: swing low, trail after +1R
16
17## Rules
18- Never hold overnight.
19- Skip days with VIX > 25.
→ Plain English rules
→ Tool-calling agent
→ Deterministic replay
→ Claude / GPT / local
02 — The workflow
From idea to live trading, in an afternoon.
One runtime, four stages. No servers to manage, no data to vendor, no brokerage plumbing to wire.
01
Write
Paste Pine. Drag in .py. Type markdown in the editor. Or start from a built-in example.
02
Backtest
20 years of history across equities, commodities, FX, and futures. Event-driven runs that model slippage and fees — results in seconds.
03
Paper test
Go live against the real tape with fake capital. Watch slippage, fees, and latency as they really happen.
04
Deploy
Connect OANDA or Pepperstone (cTrader). One click, and we run it 24/7.
3
Runtimes — Python · Pine · AI
20yr
Market history on tap
1-click
Backtest → live
Slippage
Modelled in every backtest
03 — Why BotPit
Built for traders who know better.
Reliability
Push a strategy, not a server.
No droplets to babysit, no spreadsheet reconciliation. BotPit runs, monitors and reconciles every bot for you — with kill-switches and circuit breakers built in.
Flexibility
Three runtimes, one engine.
Write bots in Python, Pine, or plain-English AI agents. The exact same backtest and live-execution engine runs all three — no rewrite when you go live.
Fidelity
Backtests that match live fills.
Every backtest models slippage and fees, so paper performance tracks your broker — not an idealised equity curve.
04 — Marketplace
Discover a bot. Or publish your own.
Every public bot on BotPit shows its source, its 5-year backtest, and its live equity curve. Fork, tune, and republish with attribution — or charge for access.