How AI is applied across API Evangelist and APIs.io. Read my AI disclosure →
API Evangelist API Evangelist
Discovery
Learnings
Guidance
Toolbox
Alignment
API Evangelist LLC

Every Talk At APIdays Paris, Mapped To The Kin Score

November 19th, 2026 · Kin Lane
Every Talk At APIdays Paris, Mapped To The Kin Score

APIdays Paris is 1 to 3 December at CNIT Forest in La Défense, part of FOST Paris. Before I go, I did what I always do with an agenda: I read every session. FOST Paris is big, with 599 sessions across 27 tracks, so I read the ten that are about APIs: Apidays, GenerationAI, GreenIO, the API Standards Conference, the Europe API Conference, APIOps, ContextCon and MCP Summit, Karate Labs, Women in API and FINOS. That is 339 sessions, and I mapped each one to the Kin Score facet it speaks to.

Same question as London: what will the people on stage tell the room an API needs to be, and is the rubric I use across the APIs.io catalog looking at the same things?

The short version. Agent Readiness leads again. It is the main subject of 65 talks and comes up in 108. What changed since London is the depth of the conversation: talks about a published contract (29), MCP servers (23) and, new this time, the quality of MCP tools (18). Having an MCP server is no longer the story; whether an agent can use it is. That lines up with the MCP Tool Quality check I added to the rubric in 0.27.0. Contract Governance and Contract Quality come next, with 25 and 24 primary talks, carried by the API Standards Conference and the APIOps day. Accountability, which I still don’t score, is the main subject of 14. And the largest group by count, 78 talks, sits outside the rubric: the GreenIO track on energy and carbon, plus a long tail of agent-engineering talks that are about building agents, not publishing APIs.

Each talk is listed once, under the facet it is mostly about, with its day and Paris start time. Many of them touch several. To filter by facet, track or day, or see every facet a talk touches, the full mapping is at talks.apievangelist.com/events/apidays-paris-2026/. Every talk and speaker links to the APIdays agenda, which is where you should go to plan your three days.

I’ll be there all week. I’m part of the APIOps Conference day on Tuesday, I have two talks on the Apidays track, Agents Inherit Your Governance and You Cannot Bolt Agent Readiness Onto Ungoverned APIs, and on Thursday I open the Europe API Conference, the EU government track I’m moderating, with a keynote on what Europe’s government APIs look like from the outside-in. Come find me.

Agent Readiness

The biggest group of talks at APIdays Paris is about whether an AI agent can drive an API without a human filling in the gaps, and this year the conversation has moved past having an MCP server to whether its tools are any good. Agent Readiness is the separate score I use to measure it: a published OpenAPI, an MCP server and the quality of its tools, machine-readable auth, idempotency, stable errors, rate-limit headers, agent skills, a well-known catalog and protected resource metadata. How the Kin Score measures it on APIs.io.

Contract Governance

Rulesets, guidelines, breaking-change detection and standards: the artifacts that constrain a contract. This is where the APIOps day, the governance talks and the OpenLint conversation live. How the Kin Score measures it.

Contract Quality

The API Standards Conference carries this facet: OpenAPI in the age of AI, JSON Schema, AsyncAPI, GraphQL and Arazzo. It is 25% of the composite, the largest facet, because everything an agent does starts from the contract. How it is measured.

Operational Transparency

Security posture, supply chains and evolving without breaking anyone: what a provider publishes about how its API behaves in production. How it is measured.

Accountability

Who answers for an organisation’s use of AI. I don’t score this yet. It is a candidate standalone layer, and the talks below are the reason it keeps coming up.

Access Clarity

What it costs and how you get in: AI spend, keys, quotas, finops and how an API is sold. How it is measured.

Developer Ergonomics

How quickly someone, or something, can get started, now that coding agents sit between the developer and the API. How it is measured.

Discoverability

Whether an API can be found and understood from machine-readable metadata alone. Most of the Europe API Conference’s government talks start here. How it is measured.

Open Source Surface

Where the product itself is open: supply-chain incidents, open collaboration in finance and open models.

Regulatory Posture

Applied only in regulated industries: auditability, sovereignty and AI governance for financial regulators. It shows up as a secondary facet in 26 talks.

Outside the rubric

The largest group by count, and two different things. The GreenIO track is about energy, carbon and water, a disclosure no API provider publishes and I don’t measure yet. The rest is agent engineering (memory, evaluation, coding factories), which is about building agents, not publishing APIs.