FreeAPI.watch

Site Statistics and Transparency

This page is auto-generated from our snapshot data on every build. We run hourly HTTP checks on every API in our catalog and publish the aggregate measurements here. The full per-API data is exposed at our public JSON API for anyone who wants to build on top of it.

Catalog overview

133
active APIs tracked
10
in the graveyard (dead or paid)
14
categories
50
recent state-change events

Performance leaders

Fastest 5 right now

  1. ipapi.co — 3 ms · geocoding
  2. CoinGecko — 5 ms · cryptocurrency
  3. TheSportsDB — 5 ms · sports
  4. PokéAPI — 7 ms · games
  5. Frankfurter — 8 ms · finance

Slowest 5 right now

  1. Nominatim (OpenStreetMap) — 853 ms · geocoding
  2. NASA Image and Video Library — 836 ms · dev-tools
  3. Open Library API — 764 ms · books
  4. DeepWiki MCP — 760 ms · mcp
  5. Visual Crossing Weather — 727 ms · weather

These are single-point latest readings, not averages. An individual slow check can place an API in the slowest list temporarily; check the per-API page for the 30-day trend.

Per-category breakdown

Category APIs tracked Alive now Avg response time
books 4 4 408 ms
cryptocurrency 15 13 132 ms
dev-tools 13 13 227 ms
entertainment 11 10 282 ms
finance 13 13 249 ms
food 4 4 109 ms
games 5 4 145 ms
geocoding 15 14 302 ms
mcp 9 9 346 ms
music 4 4 145 ms
news 13 13 227 ms
public-data 9 9 141 ms
sports 3 3 181 ms
weather 15 15 214 ms

Authentication distribution

Of the 133 active APIs we track, the authentication burden splits as follows:

Free-tier friction distribution

We score each API's free tier on a 0-1 scale based on signup friction. A no-key public API scores 1.0; an API requiring a free key with email signup scores 0.7; an API requiring credit-card details on signup scores 0.3. The split across our active catalog:

How this data is collected

Every hour, a Cloudflare Worker reads our catalog from D1, sends one HTTP request to each active API's documented health endpoint or simplest functional endpoint, and records the status code and response time. A second daily Worker rolls these into 30-day uptime percentages and average response times per API. The full methodology — including how we handle retries, timeouts, and User-Agent requirements — is documented at /methodology.

The data on this page was generated at 2026-08-09T02:57:54.966Z, the timestamp of the snapshot consumed by this build.

What's not here yet

Things we'd like to add but haven't: regional latency comparison (pinging from multiple Cloudflare edge locations), body-level response validation (catching APIs that return 200 with errors inside), and historical uptime graphs beyond 30 days. If you'd find any of these especially useful, the contact form is here.