---
type: "article"
title: "The APIs.io Demand Report #02 — the agents showed up (MCP searches +231%)"
newsletter: "APIs.io"
newsletter_handle: "apisio"
newsletter_url: "https://usecommune.com/n/apisio"
author: "Kin Lane (@kinlane)"
published: "2026-07-27T21:33:31.013Z"
canonical_url: "https://usecommune.com/n/apisio/a/CMSbR6Jp"
markdown_url: "https://usecommune.com/n/apisio/a/CMSbR6Jp.md"
chat_url: "https://usecommune.com/n/apisio/a/CMSbR6Jp/chat"
body_source: "native"
likes: 1
replies: 0
body_words: 1431
---

# The APIs.io Demand Report #02 — the agents showed up (MCP searches +231%)

*Human searches held flat. Agent searches tripled. Here's what the machines were shopping for.*

**Week of July 20 – 26, 2026 · 25,439 providers · 106,926 APIs indexed**

***

Every Monday I read what people actually typed into [apis.io](https://apis.io/) — the search box and the MCP server — as a demand signal for the API economy. Nobody else has this data, so nobody else can send you this email.

Quick housekeeping: this is Issue #02, and it lands three weeks after #01. Those weeks went into a catalog rebuild that took the index from roughly 10,000 providers to 25,439, and from 34,569 APIs to 106,926. Two things drove that. First, we started listing every provider we profile rather than only the ones with a machine-readable spec on file. Second, a full artifact refresh finally landed the specs, scopes, and security definitions for the rest of the catalog. So treat the big numbers as a re-baseline, not a one-week growth rate — and treat the demand numbers below as the real weekly signal, because those are measured the same way they were three weeks ago.

***

## 📈 The Demand Report

Raw volume went sideways: **38,015 searches, down 0.4% week-over-week.** That's the least interesting number in this email. Two others aren't:

**Unique terms rose 9.5%, from 5,553 to 6,082.** Same volume, more distinct questions. Demand is getting more specific, not bigger.

**Searches arriving over the MCP server went from 310 to 1,027 — up 231%.** That is agents, not people, and it is the fastest-moving line we have.

![This week vs last: searches flat at 38.0k, unique terms +9.5% to 6,082, MCP searches +231% to 1,027](https://fmxorstnlyoqpqdftivs.supabase.co/storage/v1/object/public/article-images/15d2e398-36ac-48cf-81d9-7b3858493d21/ed7d7e0e-e153-4147-9032-6a9e7433602c/1785187885514-qdl0mt2p.png)

Those two facts explain each other. An agent doesn't type "weather." It decomposes a task and issues a string of narrow, literal queries. When agent traffic triples, the vocabulary widens.

**And you can see it in what was searched.** The agent stack shows up as its own coherent cluster in the top 100 this week:

| Searched for | Times |
| --- | --- |
| RAG | 14 |
| MCP Server | 10 |
| Chat Completions | 10 |
| OpenAI Compatible | 9 |
| Model Gateway | 8 |
| Validated Output | 8 |
| x402 | 8 |
| Model Routing | 7 |
| AI Gateway | 6 |
| Agent Memory | 6 |

Read that list as a shopping list, because that's what it is. Nobody searching "OpenAI Compatible" or "Model Routing" wants to read about AI — they want an endpoint that speaks a protocol they've already committed to. `x402` showing up eight times is the one I'd watch: that's agent-native payments, being looked for by name, by machines.

**The rest of the top 100 is stubbornly, reassuringly boring** — and that's the other half of the story. Images (30), Weather (24), Mailgun (19), Payment Processing (14), Twilio (13), Places (12), QuickBooks (8), GitHub (8). The integration primitives that have been in demand since the first web APIs are still the volume leaders. The agentic layer is being added on top of that, not instead of it.

Underneath both, a long tail of genuinely specific enterprise asks: Carrier Billing (10), Laboratory Instruments (10), Stop Loss (10), Resource Optimization (10), Delivery Reports (9), Writebacks (9), Barcode Scanning (8), Charge Points (8), Construction Management (8), Privileged Identity Management (8), Vehicle History (8), Water Data (8), Non-Tariff Measures (8), Bills of Material (7), Utility Rates (7). Every one of those is somebody with a real integration to ship.

***

## 🕳️ The Gaps (demand we couldn't answer)

263 distinct terms came back empty this week, up marginally from 257. Tripling the catalog did *not* close the gap — because the gaps aren't categories anymore. They're names.

*A note on method: the raw unanswered list had 623 terms; 360 were dropped as noise. A scraper spent the week walking a dictionary through our search box, which is why the unfiltered list is full of things like "kielbasa" and "creme brulee." What's below is what survived.*

![The Gaps — top zero-result searches: Kimi K3, MiMo V2.5, MiniMax M3, Nano Banana and Bugbounty at 4× each, Consignment Notes, Thordata and Mydukaan at 3× each](https://fmxorstnlyoqpqdftivs.supabase.co/storage/v1/object/public/article-images/15d2e398-36ac-48cf-81d9-7b3858493d21/ed7d7e0e-e153-4147-9032-6a9e7433602c/1785187946635-6w6n1jc6.png)

**The headline gap is AI models, asked for by name.** Kimi K3, MiMo V2.5, MiniMax M3, Nano Banana — four model names, four times each, zero results. People are treating an API index as a model index. That's a genuine signal about how the two categories are collapsing, and it's a gap we own rather than one a provider can fill for us.

**The second cluster is named products nobody has submitted.** Thordata, Mydukaan, Kiwi TCMS, Altium 365, MirrorFly, GitScrum, Moysklad, NeutrinoAPI, Tangocard, Streakwave. Someone specifically wanted your API this week and got nothing. If one of those is you, that's a two-minute fix.

**The third is verticals still under-covered**, and this week they lean logistics, agriculture, and non-US: Consignment Notes, Missed Pickup, Precision Farming, Instant Payouts, POS SAFRA, Cadastro Bancário (Brazilian banking registration, asked four different ways), Techniker Krankenkasse (a German statutory health insurer). Demand is not only in English and not only in the US.

**One to flag for the security crowd:** *Bugbounty* (4×) and *Indirect Prompt Injection* (2×) both came back empty. Vulnerability-disclosure and prompt-injection-defense APIs are being looked for and are not in the index.

**If you build in one of these lanes, this is your invitation.** Add your API in about two minutes: [**apis.io/add**](https://apis.io/add) — or point your agent at the [apis.io MCP server](https://apis.io/) and let it submit for you.

***

## 🆕 New to the Index

The arrivals board — provider profiles publishing daily this week, with their real catalog surface:

- **SendGrid** — 44 APIs behind what most people think is one send-email endpoint *(Email)*
- **Zoom** — 41 APIs; a platform surface, not a marketing surface *(Communications)*
- **Adyen** — 202 APIs and 202 specs, a 1:1 ratio that's rarer than it sounds *(Payments)*
- **FactSet** — 460 APIs, 447 specs; the deepest single-provider surface in financial data *(Market Data)*
- **Shopify** — 213 APIs, 197 specs *(E-Commerce)*
- **Palo Alto Networks** — 168 APIs, 159 specs *(Security)*
- **Slack** — 115 APIs, 109 specs *(Collaboration)*

Worth noting: **"Palo Alto" was searched 9 times this week**, and the Palo Alto Networks profile publishes Saturday. Demand and supply occasionally do meet.

Full profiles publish daily on the [apis.io blog](https://apis.io/blog) through August 2.

***

## ⭐ Rated This Week: Alpaca

Every provider carries a Kin Score across six facets — discoverability, contract quality, governance, operational transparency, developer ergonomics, and commercial clarity — plus a conditional regulatory layer for providers operating under a real regime.

This week the top of the whole index, all 25,439 providers, is a company most people outside fintech have never heard of:

**Alpaca — 86.4 / 100 · Band: EXEMPLAR** 🟢 *(up from 83.7)*

The commission-free trading and market-data broker scores #1 by doing the boring things completely:

- **Discoverability: 100 · Commercial clarity: 100** — you can find it, and you can tell what it costs, without talking to anyone.
- **Developer ergonomics: 91.3** — separate, publicly maintained OpenAPI specs for Trading, Broker, Market Data, and OAuth, in the `alpacahq/alpaca-docs` GitHub repo. Not a PDF. Not a portal-only viewer.
- **Operational transparency: 86.8** — real public reliability signal.
- **Regulatory (Securities & Market Data): 100** — a perfect score on the conditional layer that only applies to providers in a regulated regime.

Where it leaves points on the table: **governance (60.5) and contract quality (69)**. The specs exist and are public, but the versioning, deprecation, and change-communication story around them is thinner than the rest of the profile. That's the cheapest ground Alpaca has left to take — and it's the same gap I see on most exemplar-band providers. Publishing the spec is now table stakes; governing it is where the remaining points live.

*The bands, for reference: exemplar 70+ · strong 60–69 · developing 45–59 · thin 30–44 · minimal 0–29.*

***

## 🤖 Reproduce this yourself

Everything above came out of public apis.io surfaces. Paste this into your agent (apis.io MCP) to pull the top-rated providers for yourself:

```javascript
find_providers(band=["exemplar"], sort="composite", limit=25)
```

Or check the demand data directly: [**apis.io/search-terms/**](https://apis.io/search-terms/) — the top 100 searches and the top 100 unanswered ones, refreshed every Monday.

***

**25,439 providers. 106,926 APIs. 38,015 searches. 263 still unanswered.** See a gap that's yours? → [**apis.io/add**](https://apis.io/add)

*The Demand Report is a weekly read of apis.io's own search + discovery signal. Forward it to someone who ships an API.*

***

## Discussion

No replies yet.