Rhumb

Agent-native tool intelligence. Discover, evaluate, and execute external tools — with trust scores, failure modes, cost-aware routing, and managed credentials.

🌐 rhumb.dev · ⚡ Quickstart · 💵 Pricing · 📊 Leaderboard · 📖 Methodology · 🔑 Trust

For agents: See llms.txt for machine-readable documentation and agent-capabilities.json for structured capability metadata.

---

Start in 30 seconds

MCP (recommended)

npx rhumb-mcp@latest

Zero config. Discovery tools work immediately — no signup, no API key.

For execution, pass your Rhumb API key:

RHUMB_API_KEY=your_key npx rhumb-mcp@latest

Get an API key →

API (read-only, no auth)

curl "https://api.rhumb.dev/v1/services/stripe/score"

All read endpoints are public.

---

What Rhumb does

Agents need external tools. Choosing the right one is hard — not because of feature lists, but because of:

• auth and signup friction
• provisioning reality vs. marketing claims
• schema instability
• failure recovery when no human is watching
• hidden costs and rate limits

Rhumb makes those constraints visible before you commit.

Rhumb Index — Discover & Evaluate

695 scored services across 50+ domains. Each gets an AN Score (0–10) measuring execution quality, access readiness, and agent autonomy support.

• find_services — search by capability, domain, or name
• an_score — full score breakdown with dimension-level detail
• get_alternatives — ranked alternatives for any service
• get_failure_modes — known failure patterns before you integrate

Rhumb Resolve — Execute

414 capabilities across 23 providers. Cost-aware routing picks the best provider for each call.

• execute_capability — call a capability through Resolve with managed auth
• resolve_capability — see ranked providers before executing
• estimate_capability — get cost estimate before committing
• Budget enforcement, credential management, and execution telemetry included

Three credential modes

Mode	How it works
BYO	Bring your own API key — Rhumb routes, you authenticate
Rhumb-managed	Rhumb holds the credential — zero setup for the agent
Agent Vault	Your key, encrypted and stored — Rhumb injects at call time

Payment paths

• API key — sign up, get a key, prepaid credits
• x402 / USDC — no signup, pay per call on-chain

---

MCP tools

rhumb-mcp exposes 17 tools:

Discovery

• find_services — search services
• an_score — score breakdown
• get_alternatives — ranked alternatives
• get_failure_modes — failure patterns
• discover_capabilities — browse capability registry
• usage_telemetry — your usage analytics

Execution

• resolve_capability — ranked providers for a capability
• execute_capability — execute through Resolve
• estimate_capability — cost estimate before execution
• credential_ceremony — set up credentials
• check_credentials — verify credential status
• routing — configure routing strategy

Billing

• budget — set spend limits
• spend — check current spend
• check_balance — prepaid balance
• get_payment_url — top-up link
• get_ledger — transaction history

---

API

Base URL: https://api.rhumb.dev/v1

Endpoint	Auth	Purpose
GET /services/{slug}/score	No	Score breakdown
GET /services/{slug}	No	Service profile + metadata
GET /services/{slug}/failures	No	Known failure modes
GET /search?q=...	No	Search services
GET /leaderboard/{category}	No	Category rankings
GET /capabilities	No	Capability registry
GET /capabilities/{id}/resolve	No	Ranked providers
POST /capabilities/{id}/execute	Yes	Execute a capability
GET /capabilities/{id}/execute/estimate	Yes	Cost estimate
GET /telemetry/provider-health	No	Provider health status
GET /telemetry/usage	Yes	Your usage analytics
GET /pricing	No	Machine-readable pricing

---

Examples

See examples/ for runnable scripts:

Example	What it shows	Auth needed?
discover-and-evaluate.py	Search → Score → Failure modes	No
resolve-and-execute.py	Resolve → Estimate → Execute	Yes
budget-aware-routing.py	Budget + cost-optimal routing	Yes
dogfood-telemetry-loop.py	Repeatable Resolve → telemetry verification loop	Yes
mcp-quickstart.md	MCP setup for Claude, Cursor, etc.	Optional

# Try discovery right now (no auth needed)
pip install httpx && python examples/discover-and-evaluate.py

---

Docs

• Agent Accessibility Guidelines — making web interfaces usable by AI agents
• AN Score Methodology — scoring dimensions, weights, and rubrics
• Architecture — scoring engine design
• API Reference — endpoint details
• Security Policy — vulnerability reporting and security architecture
• Runbooks — operational procedures
• Dogfood Loop — repeatable Resolve → telemetry validation harness

---

Repo structure

rhumb/
├── packages/
│   ├── api/         # Python API (Railway)
│   ├── astro-web/   # Public website (Vercel)
│   ├── mcp/         # MCP server (npm)
│   ├── cli/         # CLI tooling
│   └── shared/      # Shared types/constants
├── examples/        # Runnable examples
├── docs/            # Public documentation
├── scripts/         # Scoring + verification
├── artifacts/       # Score datasets
├── llms.txt         # Machine-readable docs for agents
└── agent-capabilities.json  # Structured capability manifest

---

Development

# API
cd packages/api && pip install -r requirements.txt && uvicorn app:app --reload

# MCP
cd packages/mcp && npm ci && npm run dev

# Web
cd packages/astro-web && npm ci && npm run dev

Node 24+ recommended (.nvmrc included).

---

Score disputes

Every score is disputable. If you believe a score is inaccurate:

1. Open a GitHub issue with evidence
2. Or email providers@supertrained.ai

Negative findings remain visible. Rhumb does not accept payment to change scores.

---

Links

• Website: rhumb.dev
• npm: rhumb-mcp
• MCP Registry: Rhumb on MCP Registry
• X: @pedrorhumb

License

MIT