cortex-ai-network/crypto-arbitrage-bot-automated-trading

Cortex AI is a closed source crypto trading bot distributed as a downloadable binary for Windows, Linux, and macOS. The README points users to arbitrage-bot.pro/download.php to grab the package, then to run Cortex_AI_v3.4.exe and connect exchange API keys locally. It says the keys never leave the user's machine. Despite the GitHub presence, the actual trading code is not shipped in the repository, just installation instructions and marketing copy. The bot is pitched at automated arbitrage on crypto exchanges and decentralized pools. The README lists TON, Solana, Base, BSC, and Arbitrum as supported networks, and names Binance, Bybit, Raydium, STON.fi, and Orca as venues it scans. Three trading approaches are described: latency arbitrage that tries to react to price moves before they hit public order books, funding rate arbitrage that opens opposing positions on spot and perpetual contracts, and a tracking layer the README calls Ghost Arbitrage that reads mempool data on certain chains. There is also language about local LangChain agents reading social feeds and economic calendars to adjust trading parameters around news events. The README claims sub 45ms execution loops, private RPC routing, and protection against MEV sandwich attacks via private transaction bundling. None of these claims are backed by independent benchmarks or by visible source code in the repository. System requirements listed are modest: 2 GB of RAM, 500 MB of disk, and a stable broadband connection. The product has a free download labeled v3.4 Stable plus references to a Pro plan with an Airdrop Hunter module that promises automated wallet rotation. A short developer section at the bottom shows a git clone command and a pip install line, but the meaningful logic is delivered as the pre built binary, not as readable Python in the repo. Anyone considering this should treat the claims with caution. The README is heavy on promotional language and a benchmark table comparing the bot to vague legacy Python frameworks, but the repo itself does not expose the trading logic for review, and the install path routes through an external website rather than the GitHub release page.