DEX Arbitrage Trader

Cross-DEX arbitrage bot for Base and Arbitrum. Scans live market data, detects price gaps between DEXes, and runs online dry-run (default), local paper, or live on-chain swaps.

Default mode is simulate — real QuoterV2 / Aerodrome / PancakeSwap quotes via RPC eth_call, no broadcast. Flip to live only when ready.

Modes

Mode	Behavior	Needs
paper	Local AMM math + SQLite (offline fallback)	nothing
simulate	Default — on-chain QuoterV2 round-trip (Uni/Aero/PCS); ledger updated; never sends txs	RPC URLs (ONEINCH_API_KEY optional)
live	Broadcast swaps via 1inch; hard size cap	+ WALLET_PRIVATE_KEY, funded wallet, live_enabled: true

Strategies

1. Extreme funding harvest (high risk / reward)

When HL funding is abnormally rich/cheap, take the receiving side with a hard stop. Real rates; can get stopped out. Live = HL perps.

2. HL momentum (selective)

Directional trend + funding tilt on 15m candles. Higher ROC bar to avoid fee bleed.

3. Stablecoin lending (safety sleeve)

Aave / Fluid / Compound USDC via live DefiLlama APY.

Reality check

• Funding carry is how retail-consistent PnL shows up in 2026 research (HL ↔ CEX / single-venue carry). Simulate uses live HL rates; live needs an HL wallet + hedge plan.
• Spatial DEX arb alone is not a consistent profit engine against MEV bots.

Setup

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env

Online dry-run (recommended)

1. Put RPC URLs in .env (publicnode works; dedicated Alchemy/Infura is more reliable):

   BASE_RPC_URL=https://base.publicnode.com
   ARBITRUM_RPC_URL=https://arbitrum.publicnode.com
   API_PORT=8005
   

2. Optional: ONEINCH_API_KEY for live path / aggregator mode
3. Confirm execution_mode: simulate and simulate_quote_mode: onchain in config/settings.yaml
4. Run:

   python -m src

5. Dashboard: http://localhost:8005 — badge should show SIMULATE

Live checklist (tiny size)

1. Prove repeated simulate wins (see live_require_simulate_wins)
2. Dedicated wallet with limited funds (two non-atomic legs — inventory risk)
3. Fund with USDC (+ a little ETH for gas) on Base
4. Set WALLET_PRIVATE_KEY, ONEINCH_API_KEY, execution_mode: live, live_enabled: true, keep live_max_trade_usd: 25
5. Restart and watch the first trade carefully

Configuration

Edit config/settings.yaml:

• execution_mode: paper | simulate | live
• simulate_quote_mode: onchain | pool_direct | oneinch
• live_max_trade_usd, live_enabled, spread/risk limits, DEXes

Usage

# Engine + dashboard
python -m src

# CLI monitor loop
PYTHONUNBUFFERED=1 python -m src.monitor

# Portfolio / reset
python -m src.cli status
python -m src.cli reset

API

Endpoint	Description
GET /api/portfolio	Balance, P&L, win rate
GET /api/status	Mode, 1inch/RPC readiness, risk
GET /api/trades	Trade history
GET /api/equity	Equity curve
GET /api/opportunities	Detected / rejected / executed
GET /api/activity	Live process feed

Disclaimer

Online dry-run is closer to reality than local AMM math, but still not a guarantee of live P&L. Real arb involves MEV, latency, failed txs, and stuck inventory between legs. Start tiny.