Skip to content

ivan09069/SwarmSentinel-v3

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

SwarmSentinel

Automated mean-reversion crypto trading bot. Buys RSI dips below EMA50, sells at target profit or stop-loss. Runs 24/7 on Railway, trades BTC/ETH/SOL/XRP against USDC on Coinbase.

How It Works

SwarmSentinel monitors 5-minute candles across four pairs. When RSI drops below 42 and price is below the 50-period EMA, it opens a $5 position. Exits happen at +0.8% take-profit, -0.6% stop-loss, or after 9 bars (45 min) max hold time.

Market Data → RSI Check (< 42?) → EMA50 Filter → BUY
                                                    ↓
                              SELL ← TP 0.8% / SL 0.6% / 9-bar timeout

Parameters (v5)

Parameter Value
BUY_RSI 42
SELL_RSI 70
Take Profit 0.8%
Stop Loss 0.6%
Max Hold 9 bars (45 min)
Cycle 300 seconds
Trade Size $5 USDC
Daily Limit $30
Pairs BTC-USDC, ETH-USDC, SOL-USDC, XRP-USDC
Mode Mean reversion

Architecture

v3 (this repo): LangGraph 5-node pipeline with Grok AI sentiment analysis and Kelly Criterion position sizing. Nodes: Sentinel → Analyst → Strategist → Risk Manager → Executor.

v5 (live on Railway): Streamlined single-file Python bot focused on mean-reversion. Stripped the AI sentiment layer in favor of pure technical signals for reliability.

Stack

  • Python (pure stdlib on Termux, full deps on Railway)
  • Coinbase Advanced Trade API
  • RSI + EMA50 technical indicators
  • Railway for 24/7 hosting

Related Repos

Deployment

Currently running on Railway (project ID: e20ef10a-7244-4bbd-be6d-1d9065c6dbe0). Also runs from Termux on Android at ~/swarm_sentinel/.

Roadmap

  • Deploy lmi_feed.py (Liquidity Momentum Index as Kelly multiplier)
  • Backtest OU pair scanner results (JPM/BAC best mean-reversion pair at ~8d half-life)
  • Add Telegram trade notifications via EchoForge bot

License

MIT

About

LangGraph crypto sentiment bot using Grok AI. 5-node pipeline: sentiment → whale scoring → regime detection → strategy routing.

Topics

Resources

Stars

Watchers

Forks

Packages

 
 
 

Contributors