# Rhumb — Agent Gateway for Discovery and Governed Execution > https://rhumb.dev ## What is Rhumb? Rhumb is an agent gateway for discovery and governed execution. Mission: make the internet as agent-native as possible. Rhumb Index scores external services for AI-agent compatibility. Rhumb Resolve executes supported capabilities through governed execution and routes each call to the best-fit provider for the call by first matching the supported capability path, then using AN Score, provider availability / circuit state, estimated cost, credential mode, latency proxy, and explicit policy constraints by default, while also allowing explicit provider choice when the agent wants direct control. ## System split: Index and Resolve - Rhumb Index is free discovery: score, compare, and research services. - Rhumb Resolve is per-call execution: route supported jobs through one governed surface. - Agents can say what they want done or explicitly pin the provider/tool path they want. - Index ranks the field. Resolve routes the call. ## How Resolve routes calls - AN Score is a major routing input, not the sole routing rule. - Resolve considers supported capability matching, provider availability / circuit state, estimated cost, credential mode, latency proxy, and explicit policy constraints before execution. - Agents can still pin the provider/tool path explicitly when direct control is better than automatic routing. - Discovery breadth is wider than current callable coverage, and Rhumb keeps that boundary explicit. ## Pick the right starting path - Evaluating with curl: https://rhumb.dev/quickstart#rest-path — Start with the free read endpoints. See real responses, learn the product model, and judge the trust layer before signup. - Building an agent: https://rhumb.dev/quickstart#mcp-path — Install the MCP server or use the starter prompt. Reach for capability IDs when the agent needs executable actions, not just vendor names. - Need an account and reusable key: https://rhumb.dev/auth/login — Sign up for identity, add credits, then execute with X-Rhumb-Key. Signup creates the account. Funding unlocks governed execution. Service discovery answers who to trust. Capability discovery answers what Rhumb can execute. Signup creates identity. Credits or wallet-prefund unlock governed execution. ## Rhumb Resolve Resolve is Rhumb’s governed execution layer for AI agents. Index ranks the field. Resolve routes each supported call to the best-fit provider for the call by default. AN Score is a major routing input, not the sole routing rule. Agents can also pin the supported provider path explicitly when they want direct control. Supporting authority pages: - What is Resolve?: https://rhumb.dev/resolve/what-is-resolve - Routing proof and factor explanation: https://rhumb.dev/resolve/routing - Resolve comparisons: https://rhumb.dev/resolve/compare - Key management and credential paths: https://rhumb.dev/resolve/keys - Per-call pricing explainer: https://rhumb.dev/resolve/per-call-pricing Current launchable scope: 18 callable providers, strongest in research, extraction, generation, and narrow enrichment. ## For Agents Install the MCP server for programmatic access: npx rhumb-mcp@latest MCP tools available: find_services({ query: "payment processing" }) — discover services by need get_score({ slug: "stripe" }) — detailed AN Score breakdown get_alternatives({ slug: "stripe" }) — comparable services ranked get_failure_modes({ slug: "stripe" }) — known failure patterns discover_capabilities({ domain: "communication" }) — browse capability definitions resolve_capability({ capability: "email.send", credential_mode: "byok" }) — rank providers, optionally scope to a credential mode, surface machine-readable recovery fields like recovery_hint.resolve_url, recovery_hint.credential_modes_url, and, when applicable, recovery_hint.alternate_execute_hint or recovery_hint.setup_handoff, and recover typoed capability IDs with search suggestions estimate_capability({ capability_id: "email.send", credential_mode: "rhumb_managed" }) — estimate the active execution rail, cost, and health before execution; anonymous direct system-of-record paths also preserve machine-readable execute_readiness handoffs execute_capability({ capability_id: "email.send", credential_mode: "rhumb_managed", body: { to: "user@example.com" } }) — execute through Rhumb budget({ action: "get" }) — check budget status spend({ period: "30d" }) — view spend by capability/provider check_balance() — current credit balance get_payment_url({ amount_usd: 25 }) — top up credits ## Honest current state Discovery breadth is wider than current execution breadth. Rhumb indexes 999 scored services and 435 capability definitions. Current governed execution is concentrated in 18 callable providers. ## API Base URL https://api.rhumb.dev/v1 ## Capabilities Browse all 435 capability definitions: https://rhumb.dev/capabilities - GET https://api.rhumb.dev/v1/capabilities?limit=100&offset=0 — paginated list of capability definitions - GET https://api.rhumb.dev/v1/capabilities?search=web+research — find the capability slug when you know the task but not the action ID - Each capability: { id, domain, action, description, provider_count, top_provider } - Capabilities are abstract actions (e.g. search.query, email.send) that map to concrete providers - Use find_services() when the question is which vendor should I use - Use discover_capabilities() when the question is what exact action slug should I call - Use discover_capabilities() in MCP to browse, then resolve_capability() to compare ranked providers, optionally filter by credential mode, follow recovery_hint.resolve_url, recovery_hint.credential_modes_url, and, when applicable, recovery_hint.alternate_execute_hint or recovery_hint.setup_handoff when a filtered route dead-ends, and get search suggestions when the capability ID is wrong ## API Endpoints - GET https://api.rhumb.dev/v1/pricing - machine-readable public pricing contract - GET https://api.rhumb.dev/v1/capabilities?limit=100&offset=0 — browse capability definitions - GET https://api.rhumb.dev/v1/services — list scored services - GET https://api.rhumb.dev/v1/services/{slug}/score — detailed score breakdown - GET https://api.rhumb.dev/v1/services/{slug}/failures — active failure modes - GET https://api.rhumb.dev/v1/leaderboard/{category} — ranked services by category - GET https://api.rhumb.dev/v1/search?q={query} — semantic search ## Pricing - Discovery (search, scores, browsing): Always free - Execution rails: governed API key, wallet-prefund, or x402 per-call - Provider-control modes where supported: BYOK and Agent Vault - No subscriptions, no seat fees, no minimums - Live pricing and markup terms: https://rhumb.dev/pricing - What is Resolve?: https://rhumb.dev/resolve/what-is-resolve - Routing proof and route-factor explanation: https://rhumb.dev/resolve/routing - Resolve comparisons: https://rhumb.dev/resolve/compare - Key management: https://rhumb.dev/resolve/keys - Per-call pricing explainer: https://rhumb.dev/resolve/per-call-pricing ## Categories - /leaderboard/accounting (9 services) - /leaderboard/agent-payments (4 services) - /leaderboard/ai (77 services) - /leaderboard/analytics (24 services) - /leaderboard/api-gateway (9 services) - /leaderboard/api-testing (5 services) - /leaderboard/apm (5 services) - /leaderboard/auth (25 services) - /leaderboard/authorization (5 services) - /leaderboard/background-jobs (11 services) - /leaderboard/browser-automation (15 services) - /leaderboard/cache (9 services) - /leaderboard/calendar (10 services) - /leaderboard/cdn (5 services) - /leaderboard/chat-messaging (6 services) - /leaderboard/code-execution (6 services) - /leaderboard/code-quality (7 services) - /leaderboard/communication (28 services) - /leaderboard/compliance (5 services) - /leaderboard/content-moderation (5 services) - /leaderboard/crm (10 services) - /leaderboard/customer-data-platform (5 services) - /leaderboard/customer-support (8 services) - /leaderboard/data-app (8 services) - /leaderboard/data-enrichment (6 services) - /leaderboard/data-pipeline (9 services) - /leaderboard/data-warehouse (10 services) - /leaderboard/databases (40 services) - /leaderboard/design (4 services) - /leaderboard/devops (90 services) - /leaderboard/document-processing (7 services) - /leaderboard/e-signatures (8 services) - /leaderboard/ecommerce (5 services) - /leaderboard/email (25 services) - /leaderboard/email-validation (8 services) - /leaderboard/feature-flags (11 services) - /leaderboard/fintech (6 services) - /leaderboard/forms (5 services) - /leaderboard/geolocation (5 services) - /leaderboard/headless-cms (12 services) - /leaderboard/hris (4 services) - /leaderboard/identity (4 services) - /leaderboard/identity-verification (10 services) - /leaderboard/incident-management (8 services) - /leaderboard/infrastructure-as-code (8 services) - /leaderboard/integrations-platform (7 services) - /leaderboard/iot (5 services) - /leaderboard/lifecycle-marketing (5 services) - /leaderboard/localization (7 services) - /leaderboard/logging (5 services) - /leaderboard/low-code (6 services) - /leaderboard/maps (5 services) - /leaderboard/media (8 services) - /leaderboard/message-queues (11 services) - /leaderboard/ml-experiment-tracking (10 services) - /leaderboard/monitoring (52 services) - /leaderboard/networking (10 services) - /leaderboard/notifications (6 services) - /leaderboard/payments (14 services) - /leaderboard/pdf-processing (6 services) - /leaderboard/productivity (35 services) - /leaderboard/push-notifications (4 services) - /leaderboard/real-time (6 services) - /leaderboard/reverse-etl (3 services) - /leaderboard/scheduling (4 services) - /leaderboard/screenshots (8 services) - /leaderboard/sdk-tooling (10 services) - /leaderboard/search (13 services) - /leaderboard/secrets (5 services) - /leaderboard/security-scanning (10 services) - /leaderboard/social (7 services) - /leaderboard/sso-directory (5 services) - /leaderboard/storage (12 services) - /leaderboard/subscription-billing (5 services) - /leaderboard/survey (5 services) - /leaderboard/testing (10 services) - /leaderboard/testing-automation (5 services) - /leaderboard/time-tracking (9 services) - /leaderboard/translation (5 services) - /leaderboard/vector-database (9 services) - /leaderboard/video (7 services) - /leaderboard/video-conferencing (4 services) - /leaderboard/web-scraping (6 services) - /leaderboard/web3 (5 services) - /leaderboard/webhooks (5 services) - /leaderboard/workflow-automation (10 services) - /leaderboard/workflow-orchestration (9 services) ## Scored Services (50 total) - /service/100ms — 100ms is a cloud-based video conferencing infrastructure platform providing SDKs and a REST API for embedding live video, audio, and screen sharing into applications. It offers flexible room templates, configurable track settings, recording, RTMP streaming, and a REST API for room management — designed for product teams building custom video experiences rather than using off-the-shelf solutions. [video-conferencing] - /service/ably — Real-time messaging platform with pub/sub, presence, history, push notifications, and MQTT support, built for high-scale distributed systems with global edge delivery and guaranteed message ordering. [real-time] - /service/abstract-api — Abstract API provides a suite of utility APIs — email validation, phone validation, IP geolocation, VAT validation, timezone lookup, holiday data, and more — accessible via a single account and consistent REST design. Its email validation API checks syntax, MX records, SMTP connectivity, and disposable/role-based address detection for programmatic email list hygiene. [email-validation] - /service/activecampaign — ActiveCampaign is a marketing automation and CRM platform providing a REST API for managing contacts, deals, lists, automation triggers, email campaigns, and pipelines — enabling agents to sync customer data, trigger automation sequences based on product events, and manage multi-channel marketing workflows programmatically. [email] - /service/activepieces — Activepieces is an open-source workflow automation platform with a visual builder, 100+ integrations, AI-powered step generation, approvals, and self-hosted or cloud deployment options as a Zapier/n8n alternative. [workflow-automation] - /service/acuity-scheduling — Appointment scheduling platform (part of Squarespace) with API for managing appointment types, availability, and bookings — widely used by service businesses for client scheduling. [calendar] - /service/adobe-pdf-services — Adobe's REST API for PDF operations including creation, conversion, OCR, compression, protection, and document generation — enterprise-grade with broad format support and Adobe's document processing quality. [pdf-processing] - /service/adyen — Adyen is an enterprise payments platform providing a unified REST API for card payments, digital wallets, local payment methods, payouts, and marketplace payments across 150+ countries — used by major enterprises (eBay, Uber, Spotify, McDonald's) for global payment processing. Agents use Adyen for high-volume payment processing, marketplace splits, and multi-currency payouts. [payments] - /service/aerospike — Aerospike is a high-performance distributed NoSQL database and cache optimized for real-time use cases — sub-millisecond latency at scale, with native support for key-value, document, graph, and time-series data models. Its REST Gateway and multi-language clients (Java, Python, Go, Node.js, C) enable agent workflows requiring extremely fast read/write access to large datasets at predictable latency. [cache] - /service/affinda — AI document understanding API — resume/CV parsing, invoice extraction, and custom document models; REST API with multi-format input (PDF, DOCX, image) and structured JSON output. [document-processing] - /service/africa-is-talking — African telecom API platform — provides SMS, USSD, voice, airtime, mobile payments, and IoT APIs for reaching mobile users across 20+ African markets via a single REST API, designed for developers building services for African mobile networks. [communication] - /service/agentdesk — AgentDesk provides a virtual desktop environment and tools for AI agents — a managed VM with browser, terminal, and GUI access via a Python SDK, enabling agents to operate full desktop workflows beyond browser-only automation. [browser-automation] - /service/agentops — AgentOps is an observability and debugging platform designed specifically for AI agents — providing a Python SDK for tracing multi-step agent runs, tracking LLM calls, tool invocations, and agent errors, with a web dashboard for visualizing agent execution timelines, cost/token breakdowns, and session replays. Agents use AgentOps to debug complex agent behaviors and monitor production agent performance. [ai] - /service/agentql — Web query language and API for structured data extraction from web pages, designed for agent workflows with semantic element targeting, natural language queries, and reliable extraction across site layouts. [browser-automation] - /service/agora — Agora is a real-time engagement platform providing SDKs and REST APIs for audio/video calling, live streaming, interactive broadcasting, and real-time messaging — with global infrastructure optimized for low-latency communication. Agents use Agora to manage channels, generate access tokens, and build AI-powered voice agent pipelines using the Agora Conversational AI Engine. [video-conferencing] - /service/airbrake — Error and performance monitoring platform with REST API for error aggregation, notice retrieval, and project management — one of the earliest error monitoring platforms, providing error tracking with deploy and performance monitoring. [monitoring] - /service/airbyte — Airbyte is an open-source data integration platform for syncing data from hundreds of sources to destinations — providing a connector marketplace, orchestrated syncs, incremental replication, and an API for managing connections, sources, destinations, and jobs programmatically. [data-pipeline] - /service/airtable — Airtable provides a REST API for reading and writing records in its structured database/spreadsheet hybrid — enabling agents to query tables, create and update records, manage attachments, and sync product data into Airtable bases for lightweight CRM, project tracking, and content management workflows. [productivity] - /service/linode-api — Linode (now Akamai Cloud) provides a REST API for managing cloud infrastructure — Linodes (VMs), Kubernetes (LKE), Object Storage (S3-compatible), Block Storage, Databases, and Load Balancers. Agents use the Linode API for provisioning developer-friendly cloud resources at competitive pricing, with a focus on simplicity comparable to DigitalOcean. [devops] - /service/alchemy — Blockchain infrastructure and developer platform with APIs for Ethereum, Polygon, Solana, and other chains — provides node access, enhanced APIs (NFT, token, transaction history), and webhook notifications for blockchain event monitoring. [web3] - /service/alertmanager — Prometheus Alertmanager — open-source alert routing, grouping, and notification system with REST API for managing alert rules, silences, and receiving/querying active alerts from Prometheus-based monitoring stacks. [monitoring] - /service/algolia — Hosted search API delivering sub-millisecond full-text and typo-tolerant search results; faceted filtering, geo-search, A/B testing, personalization, and a REST API with SDKs for every platform. [search] - /service/aws-bedrock — Amazon Bedrock is AWS's managed foundation model service — providing API access to models from Anthropic (Claude), Meta (Llama), Amazon (Titan), Mistral, AI21, Cohere, and Stability AI via a unified inference API, with RAG via Knowledge Bases and agent workflows via Bedrock Agents. [ai] - /service/amazon-chime-sdk — AWS real-time communications stack for audio, video, screen sharing, messaging, and media pipelines inside custom applications. [video-conferencing] - /service/cloudfront-v2 — Amazon CloudFront is AWS's global CDN — providing low-latency content delivery across 450+ PoPs, REST API for distribution management, cache invalidation, real-time access logging, and Lambda@Edge/CloudFront Functions for edge compute. Agents use CloudFront to manage distributions programmatically, invalidate cached content after deployments, and configure origin behaviors. [cdn] - /service/aws-cloudwatch — Amazon CloudWatch is AWS's native monitoring and observability service — providing a REST/SDK API for querying metrics, publishing custom metrics, managing alarms, accessing logs (CloudWatch Logs Insights), and creating dashboards. Agents use CloudWatch to query infrastructure metrics, publish custom application metrics, set programmatic alarms, and run log analytics queries. [monitoring] - /service/cognito — AWS user authentication and identity management [auth] - /service/aws-dynamodb — Amazon DynamoDB is AWS's fully managed, serverless NoSQL key-value and document database — providing a REST/SDK API for put/get/query/scan operations, DynamoDB Streams for change events, Global Tables for multi-region replication, and PartiQL (SQL-like) queries. Agents use DynamoDB for high-throughput, low-latency key-value access patterns and AWS-native state management. [databases] - /service/aws-ecs-v2 — Amazon ECS (Elastic Container Service) is AWS's container orchestration service — running Docker containers on EC2 or Fargate with task definitions, services (long-running), tasks (one-off), cluster management, and a REST/SDK API for full lifecycle control. Agents use ECS to deploy containerized services, run batch tasks, and manage container workloads in AWS environments. [devops] - /service/eks-api — Amazon EKS (Elastic Kubernetes Service) is AWS's managed Kubernetes service — providing a REST/SDK API for cluster creation, node group management, add-on management, and kubeconfig access. Agents use EKS to provision managed Kubernetes clusters on AWS, manage node groups for scaling, configure IAM Roles for Service Accounts (IRSA) for pod auth, and integrate with AWS services. [devops] - /service/aws-eventbridge — Amazon EventBridge is AWS's serverless event bus — routing events between AWS services, custom applications, and SaaS partners via rules and targets, with scheduled events (cron), event archive/replay, and a REST API for publishing and managing event routing. [background-jobs] - /service/aws-polly — Amazon Polly is AWS's managed text-to-speech service — converting text to natural-sounding speech in 60+ languages and 100+ voices, with SSML support and Neural TTS engines, via synchronous and asynchronous REST API and SDK. [communication] - /service/redshift-v2 — Amazon Redshift is AWS's fully managed cloud data warehouse — providing columnar storage, massively parallel query execution, Redshift Spectrum for querying S3 data lakes, a Data API for serverless SQL execution without JDBC/ODBC connections, and integration with the full AWS analytics ecosystem. Agents use Redshift to run large-scale SQL analytics, manage data pipelines, and query data lakes directly. [data-warehouse] - /service/aws-sagemaker — Amazon SageMaker is AWS's managed ML platform — providing infrastructure for training, tuning, and deploying ML models at scale, with built-in algorithms, managed notebooks, pipelines, feature store, model registry, real-time and batch inference endpoints, and JumpStart for pre-trained models. [ai] - /service/amazon-ses — Amazon SES (Simple Email Service) is AWS's high-volume email service — providing a REST API and SMTP interface for sending transactional and bulk emails, managing identities (domains/email addresses), handling bounces and complaints, and configuring sending rules. Agents use SES for cost-effective high-volume email at $0.10/1000 emails, integrated with AWS infrastructure. [email] - /service/amazon-sqs — Amazon SQS (Simple Queue Service) is AWS's managed message queue service — providing standard (at-least-once) and FIFO (exactly-once) queues via a REST API and SDK for decoupling microservices, buffering async tasks, and enabling reliable message delivery at any scale. [message-queues] - /service/aws-textract — Amazon Textract is AWS's managed document text extraction service — detecting and extracting text, tables, forms, key-value pairs, and signatures from scanned documents and images, via synchronous and asynchronous APIs with high accuracy on structured documents. [document-processing] - /service/aws-transcribe — Amazon Transcribe is AWS's managed speech-to-text service — converting audio to text with speaker diarization, custom vocabulary, automatic content redaction, real-time streaming transcription, and medical transcription via REST API and SDK. [communication] - /service/amazon-translate — Amazon Translate is a neural machine translation service from AWS supporting 75+ language pairs with real-time and async batch modes. It includes automatic language detection, custom terminology for domain-specific vocabulary, active custom translation, and deep integration with S3, Lambda, and Comprehend for multilingual document processing pipelines. [translation] - /service/aws — Comprehensive cloud infrastructure [devops] - /service/amberflo — Amberflo is a usage-based billing and metering infrastructure platform providing a REST API for ingesting usage events, defining meters, configuring pricing plans, computing real-time usage aggregates, and generating invoices — enabling agents and SaaS platforms to implement granular, consumption-based pricing models programmatically. [fintech] - /service/amplitude — Amplitude is a digital analytics platform — providing an HTTP API for event ingestion, a REST API for programmatic chart and cohort queries, and a Data Export API for raw event extraction. Agents use Amplitude to track product usage events, query user behavior metrics (funnels, retention, user paths), and export raw event data for custom analysis. [analytics] - /service/amplitude-cdp — Customer data platform (formerly Iteratively) with REST API for defining and enforcing event schemas, routing customer data to downstream tools, and managing identity resolution — Amplitude's data governance and routing layer. [lifecycle-marketing] - /service/ansible-v2 — Ansible is an open-source IT automation platform from Red Hat for configuration management, application deployment, and infrastructure orchestration using agentless SSH-based playbooks in YAML. [infrastructure-as-code] - /service/anthropic — Anthropic is a frontier AI platform offering Claude language models, a Messages API, tool use, structured outputs, prompt caching, batch processing, and safety-oriented controls for enterprise and developer workflows. [ai] - /service/anthropic-api — Claude model API with messages, tool use, and computer use capabilities [ai] - /service/anvil — PDF filling, e-signatures, and document workflow API — enables programmatic PDF generation from templates, embedded signature collection, and workflow automation for document-heavy business processes. [e-signatures] - /service/anyscale — Managed Ray and LLM inference platform with OpenAI-compatible API for running and fine-tuning open-source language models at scale — designed for teams deploying production AI workloads on Ray clusters. [ai] - /service/apache-kafka — Apache Kafka is the leading open-source distributed event streaming platform — providing a producer/consumer API, Admin API for topic and cluster management, and Kafka Streams for stream processing. Agents use Kafka for high-throughput event streaming, durable message queues, event sourcing, and real-time data pipeline orchestration. [message-queues] - /service/superset — Apache Superset is an open-source data visualization and business intelligence platform — providing a web-based interface for creating charts, dashboards, and SQL-based data explorations across 40+ database connectors. Its REST API enables agents to programmatically query datasets, create charts, manage dashboards, and run SQL queries against connected databases. [analytics] ## Scoring Formula AN Score = (Execution × 0.70) + (Access Readiness × 0.30) ### Execution (70% of final score, 13 dimensions) API reliability, error ergonomics, schema stability, latency distribution, idempotency support, concurrent behavior, cold-start latency, output structure, state leakage, graceful degradation, payment autonomy, governance readiness, web agent accessibility. ### Access Readiness (30% of final score, 7 dimensions) Signup autonomy, payment autonomy (access), provisioning speed, credential management, rate limit transparency, documentation quality, sandbox/test mode. ## Tier System - L4 Native (8.0–10.0): built for agents, minimal friction - L3 Ready (6.0–7.9): agents can use reliably with minor friction - L2 Developing (4.0–5.9): usable with workarounds - L1 Emerging (0.0–3.9): significant barriers to agent use ## Comparison Pages (decision surfaces for agents and operators) Index: https://rhumb.dev/compare — all comparisons in one place - https://rhumb.dev/blog/stripe-vs-square-vs-paypal — Payments: Stripe vs Square vs PayPal - https://rhumb.dev/blog/resend-vs-sendgrid-vs-postmark — Email: Resend vs SendGrid vs Postmark - https://rhumb.dev/blog/hubspot-vs-salesforce-vs-pipedrive — CRM: HubSpot vs Salesforce vs Pipedrive - https://rhumb.dev/blog/auth0-vs-clerk-vs-firebase-auth — Auth: Auth0 vs Clerk vs Firebase Auth - https://rhumb.dev/blog/posthog-vs-mixpanel-vs-amplitude — Analytics: PostHog vs Mixpanel vs Amplitude - https://rhumb.dev/blog/supabase-vs-planetscale-vs-neon — Databases: Supabase vs PlanetScale vs Neon - https://rhumb.dev/blog/twilio-vs-vonage-vs-plivo — Messaging: Twilio vs Vonage vs Plivo - https://rhumb.dev/blog/linear-vs-jira-vs-asana — Project Management: Linear vs Jira vs Asana - https://rhumb.dev/blog/anthropic-vs-openai-vs-google-ai — AI/LLM: Anthropic vs OpenAI vs Google AI Each comparison includes live AN Score data, friction maps, routing rules for agents, scenario-based recommendations, and failure mode analysis. Use comparisons when an agent needs to choose between tools in the same category. ## API Autopsies (deep technical breakdowns of individual tools) Index: https://rhumb.dev/autopsy — all autopsies in one place - https://rhumb.dev/blog/hubspot-api-autopsy — HubSpot (4.6): Rate limit traps, cross-hub API inconsistency, OAuth maze, no idempotency - https://rhumb.dev/blog/salesforce-api-autopsy — Salesforce (4.8): SOQL barrier, governor limits, sandbox/production split, sales-gated onboarding - https://rhumb.dev/blog/twilio-api-autopsy — Twilio (8.0): What agent-native almost looks like. Simple auth, idempotency, error codes that teach - https://rhumb.dev/blog/shopify-api-autopsy — Shopify (7.8): GraphQL-first bet, query cost budgets, forced version migration Each autopsy dissects a single tool's failure modes, score anatomy, integration cost, and constructive recommendations. Use autopsies when an agent needs to understand the specific friction of a tool before integrating. ## Getting Started - https://rhumb.dev/quickstart — Make your first API call in 30 seconds. Copy-paste examples for curl, JS, Python, MCP. No signup required for read endpoints. - https://rhumb.dev/payments/agent — x402 agent payments: HTTP 402 flow, USDC on Base, wallet setup, and when to use on-chain per-call authorization - https://rhumb.dev/blog/how-agents-actually-pay-x402-dogfood — x402 seller dogfood report: 5 compatibility mismatches, authorization proof vs tx_hash gap, and why structured 422 errors beat infinite 402 loops - https://rhumb.dev/blog/getting-started-mcp — MCP install guide, framework setup (Claude Desktop, Cursor, direct stdio), 3 workflow walkthroughs, credential paths explained - https://rhumb.dev/blog/securing-keys-for-agents — How to secure API keys for agent use: three credential paths (Rhumb-managed, BYOK, Agent Vault), storage hierarchy, honest threat modeling, and where x402 fits as a payment rail - MCP tools reference with examples for all 21 tools - Three credential paths: Rhumb-managed, BYOK, Agent Vault - End-to-end workflow example: find → evaluate → resolve → execute ## Trust and disputes - Trust overview: https://rhumb.dev/trust - Methodology: https://rhumb.dev/methodology - Current self-assessment: https://rhumb.dev/blog/we-scored-ourselves - Historical baseline: https://rhumb.dev/blog/self-score - Provider guide and dispute process: https://rhumb.dev/providers - Public dispute template: https://github.com/supertrained/rhumb/issues/new?template=score-dispute.md - Public dispute log: https://github.com/supertrained/rhumb/issues?q=is%3Aissue+%22Score+dispute%3A%22 - Private disputes: mailto:providers@supertrained.ai?subject=Score%20Dispute - Dispute response target: 5 business days ## Extended Context - Glossary: https://rhumb.dev/glossary - About the team: https://rhumb.dev/about - Extended version: https://rhumb.dev/llms-full.txt ## Links - Website: https://rhumb.dev - GitHub: https://github.com/supertrained/rhumb - MCP Server: npx rhumb-mcp@latest - Public agent capabilities: https://rhumb.dev/.well-known/agent-capabilities.json