Imagine: the mempool is congested, ETH price on Uniswap and Curve diverges by 0.3% in a second. Your MEV bot must assemble a bundle and send it to Flashbots before dozens of competitors. A delay of more than 10 ms and the profit goes to someone else. We build infrastructure that wins this race: from simulation to bundle submission. Average profit per arbitrage trade is 0.05–0.15% — on a $100k volume that's $50–150 per second. Our bots specialize in DEX arbitrage and position liquidations on Aave, Compound, and other protocols. The task: detect a profitable opportunity, simulate its execution, and send the bundle through Flashbots faster than competitors. Over 5 years, we have launched more than 50 such systems. Each bot is customized to specific strategies and client infrastructure.
How Does an MEV Bot Work?
The bot consists of an off-chain detector, an on-chain executor, and a relay network. The detector analyzes the mempool and blockchain state, the executor performs swaps via smart contracts, and the relay sends bundles directly to builders, bypassing the public mempool.
How We Detect Arbitrage Opportunities
Arbitrage between DEXes is the primary income source. We monitor prices on Uniswap, Curve, Balancer. When price divergence occurs, we compute the optimal swap size using a mathematical model. For Uniswap v2, we use the analytical formula:
optimal_amount_in = sqrt(reserve0 * reserve1 * 997 / 1000) - reserve0 For concentrated liquidity v3, we need step-by-step calculation via QuoterV2. Flash loan fees from Aave v3 are 0.05% — all calculations account for this cost.
Why Transaction Simulation Is Critical
Before sending a bundle, you must know the exact profit after gas and bribe. We use our own simulation engine built on revm (Rust EVM implementation). It executes transactions on a forked copy of mainnet in 1–3 ms, 10–20 times faster than standard eth_call. This advantage allows us to win front-running in competitive situations. The engine uses the Shanghai revision with PUSH0 support. We fork mainnet via eth_createAccessList and execute transactions in an isolated environment. The concept of MEV is described in Ethereum Foundation research.
| Strategy | Average Profit Margin (3 months) | Frequency per day |
|---|---|---|
| Uniswap v2 arbitrage | 0.05–0.15% | 200–500 |
| v3 arbitrage | 0.02–0.10% | 300–800 |
| Aave liquidations | 5–10% of liquidated position | 10–50 |
| Bundle Submission Method | Latency | Failure Risk | Setup Complexity |
|---|---|---|---|
| Direct to builder | <10 ms | Low | High |
| Flashbots relay | 20–50 ms | Medium | Medium |
| Public mempool | >100 ms | High | Low |
Problems We Solve
High latency. Public RPCs add 50–200 ms. For liquidations this is acceptable, but for arbitrage it's critical. We set up our own geth node with WebSocket subscription to txpool.
Gas optimization. Every extra wei of gas reduces profit. We compile contracts with maximum optimization in Foundry and use minimal executor code.
Competition. Dozens of bots react to the same opportunity simultaneously. The winner is the one who offers a higher bribe to the builder. We analyze MEV-Boost auction history and select the optimal bribe.
Unlike sandwich attacks, our arbitrage and liquidation strategies do not harm regular users. Sandwich attacks manipulate prices around a user's transaction, causing direct damage. We have consciously avoided such practices in favor of legitimate MEV extraction methods.
How We Develop an MEV Bot
The development process is divided into several stages, each agreed with the client.
Strategy and Market Analysis
Determine which MEV sources are priority (arbitrage, liquidations, triangulation). Assess the capital required to start and expected returns.
Architecture Design
Design the off-chain detector and on-chain executor. Choose language: Python for prototype, Rust for production. Determine which DEXes and protocols to support.
Implementation
Write detector, executor, and relay code. Integrate the revm-based simulator. Perform unit testing and fork tests on mainnet.
Testing and Optimization
Run on testnet with real data. Optimize gas, set profit thresholds, test with various builders.
Deployment and Monitoring
Deploy the node, contracts, set up logging in Graylog. Provide 24/7 monitoring via configured alerts.
Bot Architecture
Off-chain Detector
Written in Python for prototype, then Rust for production. Subscribes to mempool via WebSocket, parses transactions, calculates potential profit. For liquidations, it uses The Graph to monitor position health factors.
On-chain Executor
Simple Solidity smart contract. Accepts token address, amount, router. Performs swap via Uniswap or Curve router. Checks that profit is above a minimum threshold (minProfit). Protects against reentrancy with nonReentrant from OpenZeppelin.
Relay Network
Bundle submission is done via Flashbots MEV-Boost directly to builders (beaverbuild, rsync, titan). We do not rely on the public mempool.
What You Get After Development
Full source code of the bot (detector, executor, relay). Documentation for installation, configuration, and monitoring. Access to git repository with change history. Team training on bot management and log interpretation. 2 months of support with guaranteed stable operation. Option to extend support on individual terms.
Our Expertise
We have specialized in MEV for over 5 years. Developed and launched bots for 50+ clients. We ensure transparency: you get full access to code and operation logs. We provide a warranty on stable operation for the first 2 months. Gas savings up to 20% through executor optimization. Potential profit from 0.05% per trade, average daily profit from $500 to $5000.
Contact us for a consultation — we will assess your strategy and propose the optimal solution. If you want an MEV bot optimized for your infrastructure, get in touch for a detailed audit. Order turnkey MEV bot development and start earning on arbitrage within 4 weeks.