Public APIs Directory Scraper — 1400+ Free APIs, No Key Required

Scrape the entire public-apis directory — the largest community-maintained list of free and public APIs on the internet — into a clean, structured dataset. One row per API with its name, description, category, auth type, HTTPS support, CORS policy and documentation link.

No API key, no login, no proxy. Just run it and get ~1,600 machine-readable API records in under a minute.

---

What you get

Field	Description
name	API name, e.g. Open-Meteo
description	Short description of what the API does (capped at 1000 chars)
category	Directory category, e.g. Weather, Cryptocurrency, Machine Learning
auth	Auth type — normalized: None, apiKey, OAuth, X-Mashape-Key, User-Agent
requires_auth	Boolean shortcut — false means you can call it with zero credentials
https	Boolean — whether the API is served over HTTPS
cors	CORS policy — Yes, No or Unknown
link	Documentation / homepage URL
source	Which upstream source the record came from (github-markdown or publicapis.org)
scraped_at	ISO-8601 UTC timestamp of the run

Every field is nullable, so downstream schemas never break when the community changes a column.

Sample output

{
  "name": "Open-Meteo",
  "description": "Global weather forecast API for non-commercial use",
  "category": "Weather",
  "auth": "None",
  "requires_auth": false,
  "https": true,
  "cors": "Yes",
  "link": "https://open-meteo.com/",
  "source": "github-markdown",
  "scraped_at": "2026-07-28T14:13:31.442471+00:00"
}

---

Input

Field	Type	Default	Description
mode	select	apis	What to scrape. apis returns all directory entries.
category	string	—	Exact category filter (case-insensitive), e.g. Weather.
authFilter	string	—	Exact auth filter (case-insensitive). Use None for zero-auth APIs.
httpsOnly	boolean	false	Only return APIs served over HTTPS.
searchTerm	string	—	Case-insensitive substring match on name + description.
maxItems	integer	1000	Max records to push (hard cap 5000).

Example: every free, no-auth, HTTPS weather API

{
  "mode": "apis",
  "category": "Weather",
  "authFilter": "None",
  "httpsOnly": true,
  "maxItems": 100
}

Example: everything

{ "mode": "apis", "maxItems": 5000 }

---

Data source

The actor tries two upstream sources, in order:

1. https://api.publicapis.org/entries — the classic JSON endpoint. This host has been unreliable for a long time (DNS resolution frequently fails entirely), so it is only attempted first and silently skipped when unavailable.
2. https://raw.githubusercontent.com/public-apis/public-apis/master/README.md — the canonical markdown README of the public-apis/public-apis repository. This is the actual source of truth that the JSON API was built from, and it is what the actor uses in practice today.

At the time of the last verification run, source 2 was used and returned 1,613 APIs across 51 categories, all with a name, category, auth type and link. The source field on every record tells you exactly where the data came from, so you always know what you are looking at.

Markdown tables are parsed by tracking the current ### Category heading and reading each | [Name](link) | Description | Auth | HTTPS | CORS | row. Promotional/sponsor tables at the top of the README are filtered out, backticks and inline HTML are stripped, and No/empty auth values are normalized to None.

---

Use cases

• API discovery — find a free API for weather, sports, crypto, government data or anything else, filtered by whether it actually needs a key.
• Developer tooling — power an internal API catalog, a VS Code extension, a Slack bot or an LLM tool registry with fresh directory data.
• Integration research — audit which categories have HTTPS + CORS coverage before committing to a vendor, or shortlist candidates for a client integration.
• Hackathons — pull the whole no-auth, CORS-enabled subset in one run and hand your team a ready-made list of APIs they can call straight from the browser.

---

Pricing

Roughly $0.50 per 1,000 results. A full directory pull (~1,600 APIs) costs well under a dollar and finishes in seconds — the actor makes a single HTTP request and does all the work locally.

---

Related actors

• GitHub Trending Scraper
• npms Search Scraper
• OpenRouter Models Scraper
• pkg.go.dev Scraper

---

Notes

• The actor fails loudly with a RuntimeError if zero results are produced — either because both upstream sources are down, or because your filters matched nothing. Loosen the filters and re-run.
• Data is community-maintained. Some entries link to APIs that have since gone offline or changed their auth model; always verify before shipping to production.
• Output ordering follows the directory: alphabetical by category, then alphabetical by API name.