What is Rhumb Resolve? Resolve is governed execution for AI agents. Index ranks the field. Resolve first matches the supported capability path, then routes supported calls using AN Score, provider availability / circuit state, estimated cost, credential mode, latency proxy, and explicit policy constraints while preserving explicit provider pinning.

Resolve authority

What is Resolve?

Rhumb Resolve is the governed execution layer for AI agents. It lets an agent ask for the job, estimate the route, execute through a controlled rail, and explain which provider path was used.

The mental model is deliberately simple: Index ranks. Resolve routes. Index gives the agent service intelligence. Resolve turns supported work into auditable calls.

The wedge is One key, many superpowers. Start with managed capabilities first; bring your systems only when needed.

Discover

Index tells the agent who is worth trusting.

Rhumb Index is the public intelligence layer: scores, failures, comparisons, and capability discovery across the service catalog.

Resolve

Resolve turns intent into a governed call.

The agent can ask for a capability and let Resolve choose the best-fit supported provider, or pin a supported provider path directly.

Control

The operator keeps the boundary explicit.

Credential path, budget, provider choice, and policy constraints are surfaced before execution instead of hidden behind magic routing.

Execution lifecycle

Resolve starts before the call.

The product is not just the HTTP request. The product is the route decision, cost estimate, credential boundary, execution receipt, and operator controls around repeated agent traffic.

01

Find the service or capability

Use Index when you need vendor intelligence. Use capability discovery when you know the job but not the provider path. The bounded web-search quickstart starts with `search.query` for exactly this reason.

02

Estimate the route

Check readiness, credential mode, price, provider availability, cost, credentials, and the explainable route before the agent spends money or touches a system.

03

Execute with receipts

Call through governed API key, wallet-prefund, x402, BYOK, or Agent Vault depending on the job and trust boundary.

04

Inspect and constrain

Use spend, budgets, route explanations, and provider attribution so repeat agent traffic stays operable.

What it is not

No magic layer. No overclaim.

  • Not a generic connector directory with a nicer UI.
  • Not a claim that every indexed service is executable today.
  • Not a blind top-score router that pretends quality, cost, and credentials are the same thing.
  • Not a demand that agents give up explicit provider control.

Current boundary

Launchable where the route is real.

Current launchable scope: research, extraction, generation, and narrow enrichment — not general business-agent automation. Not every service or capability in the index is executable through Rhumb today. Discovery breadth is wider than current callable coverage.

Browse executable capabilities →