We build open-source AI systems, developer tooling, and protocol intelligence for the XRP Ledger ecosystem — from MEV research to production-grade SDKs.
Every tool we ship is grounded in published research. We don't build features — we solve problems that matter on-chain.
Peer-reviewed papers on MEV dynamics, consensus prediction, anomaly detection, and LLM agents operating autonomously on-chain.
Browse papersTyped SDKs, streaming clients, DeFi primitives, and agent frameworks — production-ready packages built for the modern TypeScript ecosystem.
View toolingReal-time risk scoring, validator reputation systems, and federated learning pipelines that surface signal from every ledger close.
Learn moreOur SDK handles XRPL complexity so you can focus on your application. Typed, documented, and battle-tested on mainnet.
Peer-reviewed work on AI, MEV, and protocol design on the XRP Ledger.
We study maximal extractable value (MEV) dynamics on the XRP Ledger, identifying frontrunning patterns and sandwich attacks in the XRPL DEX. We quantify $4.2M in extractable value and propose AI-driven countermeasures.
XRPFormer, a transformer-based model predicting XRPL consensus round latency with 94.3% accuracy up to 12 seconds in advance using validator topology and mempool depth signals.
A framework for deploying LLM agents capable of executing multi-step financial operations on XRPL. Our best model achieves 81% task completion across 240 real-world tasks with no human supervision.