Slack indexes at apis.io/providers/slack/ with 115 APIs and 109 API specifications, a Kin Score of 73.1 (exemplar), and an Agent Readiness score of 72.1 in the agent-native band.
Composite 73.1, agent readiness 72.1 — nearly identical. That balance is rare and it is the most interesting number on the page.
What’s in the surface
Slack decomposes into 32 named collections, mirrored across Postman and open sets. The boundaries are unusually clean because they follow the product’s own nouns:
- Messaging core — Chat, Conversations, Files, Reactions, Pins, Stars, Bookmarks, Emoji, Reminders
- Identity and access — Auth, OAuth, OpenID Connect, Users, User Groups, Team, Bots
- App platform — Apps, Functions, Workflows, Views, Dialog
- Newer surfaces — Assistant, Canvases, Lists
- Administration — Admin, Migration, DND, Calls, Search, Tests
- Transport — Web API, RTM
Thirty-two boundaries across 115 APIs is a finer decomposition than Shopify’s five across 213 — and it is why an agent can navigate this catalog. Each surface does one thing and is named for it.
What’s interesting
Assistant, Functions, and Workflows together describe a platform that stopped being a chat app. Functions and Workflows are automation primitives — Slack executing logic rather than relaying messages. Assistant is the AI surface. Those three are where the product is going, and they are specified rather than previewed.
RTM alongside Web API is the honest legacy note. Real Time Messaging is the older websocket transport; the Web API is the modern surface. Slack kept RTM described rather than quietly deprecating it into an undocumented state — which is more than most providers do with a superseded transport.
Canvases and Lists are the document-shaped surfaces. A collaboration platform accumulating structured content types is a platform absorbing the adjacent tools its customers were paying for separately.
Why the agent-readiness balance matters
Compare the spreads across this week’s profiles:
| Provider | Composite | Agent Readiness | Gap |
|---|---|---|---|
| FactSet | 73.5 | 76.0 | +2.5 |
| Slack | 73.1 | 72.1 | −1.0 |
| Adyen | 77.8 | 73.1 | −4.7 |
| Palo Alto Networks | 75.3 | 65.4 | −9.9 |
| Shopify | 68.7 | 53.8 | −14.9 |
Slack’s near-zero gap says the surface is as consumable by a machine as it is by a person. That follows directly from the fine-grained decomposition: an agent that wants to post a message reaches for Chat, not for a general-purpose endpoint whose behaviour depends on which fields it happens to set.
Takeaway
115 APIs, 109 specifications, 32 clean boundaries, and the tightest composite-to-agent-readiness balance in this week’s sweep. Slack’s score comes from naming things precisely — which turns out to be the same discipline that makes an API usable by software.
Browse it at apis.io/providers/slack/.