Public docs, in plain English and precise detail. No black box.

Technical documentation hub for the Loxley AI white-label Telegram trading bot: auto-buy execution flow, trailing stop-loss engine, Uniswap V4 hook safety scanner, snipe detection, multi-bot sharded architecture and operator fee engine.

This is the public technical documentation of the Loxley AI white-label Telegram trading bot for Robinhood Chain. We publish how the system works because operators should understand the machinery that carries their brand — and because transparency beats marketing adjectives. Trading is risky; nothing here is financial advice.

Core engine documentation

• Auto-buy execution flow — the complete path from pair-creation event to executed buy: detection, quoting, the safety gate, simulation and broadcast, including what each step costs in latency and why.
• Trailing stop-loss and sell engine — how positions are tracked, how trailing stops and take-profit ladders are evaluated, and why sells are never blocked by the safety scanner.
• Uniswap V4 hook safety scanner — how the scanner inspects V4 pool hooks, what patterns are flagged, and how hook-aware quoting differs from naive V2-style routing.
• Snipe and new-pair detection — creation-transaction anchoring, deployer and LP-funder analysis, and snipe-concentration scoring for fresh launches.
• Multi-bot sharded architecture — how a coordinator plus multiple pusher/receiver bot tokens scale one brand to thousands of users without hitting Telegram limits.
• Operator fee engine — plan tiers, per-trade fee accounting, revenue share, the fee wallet, and Stripe subscription lifecycle.

Reading order for new operators

1. Start with the operator onboarding runbook for the 48-hour deployment sequence.
2. Read the auto-buy flow to understand what happens in the seconds after a new token appears.
3. Read the hook scanner doc — it is the differentiating capability on Robinhood Chain.
4. Finish with the fee engine doc so your revenue math matches ours.

Design principles (the short version)

• On-chain truth over event guesses: the analyzer verifies supply, locks and taxes against live contract state rather than trusting event logs alone.
• Users decide on manual buys: the scanner informs with clear warnings; hard blocks apply to unattended automation only.
• Sells are sacred: no subsystem is allowed to block or delay an exit. Aggregator sell simulations lead over internal estimates.
• Degraded is never cached: when an RPC or scanner call fails, the system reports degraded state instead of silently reusing stale verdicts.
• Kill switches everywhere: every automated lane can be halted per operator without redeployment.

Start reading: auto-buy flow →

Built for operators, not slippage goblins. Talk to us →