Abstract

Raphael Sentinel is an on-chain intelligence engine for Solana that focuses on one thing: detecting asymmetric risk before retail gets farmed. Instead of generic “token scanners”, Raphael Sentinel runs a dual-path analysis pipeline that understands two completely different threat models:

1. Standard SPL tokens – where the main risks are LP security, mint/freeze authorities, holder concentration, wash trading and bot-driven manipulation.

2. Launchpad tokens (Pump.fun, Lets Bonk and similar) – where LP is bonded and cannot be rugged, but dev-controlled multiwallet swarms, pre-funded entries and coordinated exit dumps destroy late buyers.

Given a token contract address, Raphael Sentinel automatically detects the token type and routes it through the appropriate analysis path. For standard SPL tokens, the engine inspects LP locks and providers, authorities, holder distribution, wash-trading patterns, bot activity and wallet quality to compute a transparent risk score (0–100) with a full breakdown of red flags.

For launchpad tokens, Raphael Sentinel builds a funding graph from early buyers, identifies dev clusters and pre-funded multiwallet swarms, tracks real-time exit patterns and measures exit asymmetry between retail and the dev cluster. The result is a “farm intelligence” layer that can classify patterns such as instant farms, slow bleeds and classical pump-and-dump behavior with an explicit confidence level.

The platform is delivered as a Web3-native SaaS web app, with a FREE tier (static snapshots, limited depth) and a PRO tier that unlocks dynamic, real-time monitoring, clustering, dev history, exit tracking and alerting. All computations are backed by FastAPI microservices, Celery workers, Redis caching and PostgreSQL storage, integrated with Helius, Solana RPC and Birdeye for on-chain and market data.

Raphael Sentinel is designed from day one to integrate its own Solana token as a native payment, discount and burn asset, enabling Web3 wallet-based subscriptions and deflationary tokenomics without sacrificing sustainability of the underlying intelligence engine.