Rhumb serves as the essential infrastructure layer for AI agents, enabling them to reliably discover, access, and trust external tools and APIs. It addresses the unique challenges agents face when interacting with APIs not originally designed for autonomous use, such as authentication friction, schema instability, and unclear failure modes. By offering detailed scoring across execution quality and access readiness, Rhumb empowers agents and operators to make informed decisions, ensuring robust tool integration, predictable performance, and adherence to budgets before execution.
主な機能
01212 scored services across 419 capabilities
02Agent-Native (AN) Score for tool reliability and access readiness
03Cost-aware routing with pre-execution budget enforcement
04Support for three credential modes: BYO, Rhumb-managed, and Agent Vault
05Zero-signup per-call payment via x402 / USDC
060 GitHub stars
