The facets
Each card pairs what the speakers are saying with what the Kin Score checks. Click a card to filter the agenda below. Counts include every talk that touches a facet; "primary" is the facet a talk is mostly about.
Composite facet
Conditional facet (applies only where relevant)
Standalone score
Not measured yet
Agent Readiness45 talks · 32 primary
What you will hearBill Doerrfeld on how agents consume APIs differently, Curity on agent identity and the OAuth toolkit, Spendesk on MCP for regulated finance APIs, CVS Health on governing each agent action by its blast radius, Amadeus on the 200 OK that hides a broken procedure, WEX on agents paying per call with HTTP 402.
What the Kin Score measuresA separate score from the composite. It checks whether an agent can drive the API without a human filling in the gaps: a published OpenAPI, an MCP server, machine-readable auth, idempotency, stable errors, rate-limit headers, agent skills, a well-known catalog, protected resource metadata, and bound vs bearer credentials.
How the Kin Score measures Agent Readiness on APIs.io →
Contract Quality15 talks · 5 primary
What you will hearMSD's benchmark showing badly designed APIs cost models 30-77% more tokens, adorsys on schema-based contract testing against drift, WunderGraph on consumer-first GraphQL federation, Erik Wilde on moving repeated processes out of prompts and into workflows, Aklivity on streaming APIs as contracts.
What the Kin Score measuresThe depth of the OpenAPI, AsyncAPI, GraphQL and JSON Schema a provider actually publishes: schemas, examples, descriptions, versions. 25% of the composite, its largest facet.
How the Kin Score measures it on APIs.io →
Contract Governance20 talks · 9 primary
What you will hearDaniel Kocot on spec-driven design with a shared constitution, API Peak on governance in CI/CD, Lorna Mitchell on layered standards for agents, LSEG on measuring organisational coherence, Axway, Frends and Kong on governing the agentic era.
What the Kin Score measuresThe artifacts that constrain a contract: published rulesets, vocabulary, declared conformance, overlays. It reads what a ruleset declares, not the result of running it.
How the Kin Score measures it on APIs.io →
Operational Transparency25 talks · 9 primary
What you will hearAPIContext on the third of endpoints that degrade while the status page stays green, Apica on agentic-ready observability, the OWASP API Top 10, Akamai on AI-automated attacks, and the gateway talks from ICA Gruppen, Solo.io, Polar and Aman.
What the Kin Score measuresWhether a provider publishes how its API behaves in production: rate limits, a status page, a change log, a deprecation policy, a security disclosure path.
How the Kin Score measures it on APIs.io →
Regulatory Posture10 talks · 2 primary
What you will hearFAPI 2.0 and sender-constrained tokens for German public-sector APIs, NIIS on trusted research environments and data spaces, Spendesk on open-banking discipline for MCP, CVS Health on agents in healthcare, dormakaba on APIs that control physical access.
What the Kin Score measuresApplied only in regulated industries: consent-scoped authorization, a published security and disclosure posture, terms and privacy as a legal basis, and conformance to the industry's data standard.
How the Kin Score measures it on APIs.io →
Access Clarity9 talks · 2 primary
What you will hearZuplo on MCP monetization, WEX on pay-per-use APIs with no sign-up and no key, PagerDuty on going from two scopes to a hundred, Curity on identity as a business enabler, Open Empower on quotas and cost controls.
What the Kin Score measuresWhat it costs, what you are allowed to do, and how you get in: plans, pricing, sign-up, terms, privacy, a finops mapping, trust and compliance.
How the Kin Score measures it on APIs.io →
Developer Ergonomics6 talks · 1 primary
What you will hearWillow on choosing between API, MCP, CLI and browser, pentacor on an agent with a raw OpenAPI versus an MCP server, Spendesk's three years from REST to MCP, ZEISS on AI inside a platform team.
What the Kin Score measuresHow quickly someone can get started: SDKs, CLI, portal, getting-started guide, sandbox or console, MCP server, clear auth documentation.
How the Kin Score measures it on APIs.io →
Discoverability10 talks · 2 primary
What you will hearSAP on Open Resource Discovery, TeejLab on finding the MCP servers you already have, The Value Hub on knowing which consumers a schema change breaks, Akamai on shadow APIs.
What the Kin Score measuresWhether the API can be found and understood from machine-readable metadata alone: apis.json completeness, tags, identity signals.
How the Kin Score measures it on APIs.io →
Open Source Surface2 talks · 0 primary
What you will hearSolo.io's open-source agentgateway, and Open Resource Discovery as a Linux Foundation standard.
What the Kin Score measuresApplied only where the product itself is open source: a vulnerability disclosure path, a contribution route, a release history, a code of conduct.
How the Kin Score measures it on APIs.io →
Create-or-Update Ergonomics2 talks · 0 primary
What you will hearAmadeus on execute, verify, confirm for irreversible writes, and Adobe on the agent harness quietly doing your retries.
What the Kin Score measuresApplied only where an API accepts writes: is there a create-or-update, and is a write safe to call twice?
How the Kin Score measures it on APIs.io →
Accountability12 talks · 1 primary
What you will hearAPIsFirst on the psychology of being AI-ready, CVS Health on keeping each agent decision accountable and explainable, WSO2 asking who is accountable at each stage of an agent's growth, i-val on multi-agent systems as a game of telephone.
What the Kin Score measuresStandalone and not yet scored. Who answers for an organisation's use of AI: a published policy that resolves, versioned, with a named owner, an accountable officer and an impact assessment.
Outside the rubric1 talks · 0 primary
What you will hearAxa France's API Green Score: energy and carbon per endpoint, measured and enforced in CI.
What the Kin Score measuresNothing yet. The Kin Score reads the Spectral rules and pagination that talk automates, but has no facet for energy or carbon disclosure.