OpenAPI Directory & API Specs Scraper (APIs.guru) — 2,500+ APIs, No API Key

Scrape the APIs.guru OpenAPI directory — the largest open, machine-readable index of public REST APIs — into a clean dataset. One row per API with the OpenAPI/Swagger spec URL (JSON + YAML), all published versions, the provider, categories, contact, license, terms, logo and — optionally — the actual path and operation counts parsed from the spec itself.

No API key. No login. No proxy required.

Why this actor

APIs.guru is the "Wikipedia for Web APIs": ~2,529 APIs, ~3,992 specs and ~108,000 endpoints, kept in sync with the upstream providers (AWS, Azure, Google, Twilio, Adyen, GitHub, eBay, Stripe-likes, government open data and hundreds more). This actor turns that directory into a flat, spreadsheet-friendly dataset you can filter, join and feed into code generators.

How is this different from public-apis-scraper?

	openapi-directory-scraper (this actor)	public-apis-scraper
Source	APIs.guru /v2/list.json — a machine-readable OpenAPI directory	The public-apis curated markdown list on GitHub
Per-API data	Actual OpenAPI/Swagger spec URLs (JSON + YAML), spec version (2.0 / 3.x), every published version, spec origin URL, license, terms, contact, logo	Name, description, auth type, HTTPS/CORS flag, doc link
Machine usable	Yes — spec URLs can be fed straight into codegen / linting / mocking	No — human-curated links only
Endpoint counts	Optional: downloads each spec and counts paths + operations	Not available
Scale	~2,500 APIs, ~4,000 specs, 677 providers	~1,400 curated entries

They complement each other: use public-apis-scraper to discover free APIs a human curated, use this actor when you need the actual specs to build against.

Use cases

• API discovery — find every public API in a category (cloud, financial, payment, ecommerce, open_data, messaging, machine_learning) with a real, downloadable spec.
• Integration research — compare providers, check which spec version (Swagger 2.0 vs OpenAPI 3.x) an API ships, and grab licensing/terms links before you commit.
• SDK generation pipelines — feed spec_url into openapi-generator, Speakeasy, Fern, Kiota or Prism to auto-build clients, mocks and test suites for hundreds of APIs at once.
• Competitive API analysis — measure API surface size (path_count, operation_count) across competitors, track how many services a provider publishes, and monitor updated timestamps for API churn.

Input

Field	Type	Default	Description
mode	select	apis	apis = the full OpenAPI directory (one item per API). providers = the provider list with API counts.
providerFilter	string	–	Keep one provider only, e.g. azure.com, googleapis.com, amazonaws.com, twilio.com.
categoryFilter	string	–	Keep one APIs.guru category, e.g. cloud, financial, developer_tools.
searchQuery	string	–	Case-insensitive substring over API id, title and description, e.g. payment, weather.
countEndpoints	boolean	false	Download each preferred spec and count paths + operations. One extra request per API, specs can be multi-MB → much slower.
maxItems	integer	500	Cap on pushed items (max 5,000).

Filters combine with AND.

Example input

{
  "mode": "apis",
  "categoryFilter": "financial",
  "searchQuery": "payment",
  "countEndpoints": false,
  "maxItems": 200
}

Output

{
  "type": "api",
  "api_id": "1forge.com",
  "provider": "1forge.com",
  "service": null,
  "title": "1Forge Finance APIs",
  "description": "Stock and Forex Data and Realtime Quotes",
  "preferred_version": "0.0.1",
  "version_count": 1,
  "all_versions": ["0.0.1"],
  "openapi_version": "2.0",
  "spec_url": "https://api.apis.guru/v2/specs/1forge.com/0.0.1/swagger.json",
  "spec_yaml_url": "https://api.apis.guru/v2/specs/1forge.com/0.0.1/swagger.yaml",
  "origin_url": "http://1forge.com/openapi.json",
  "origin_format": "swagger",
  "categories": ["financial"],
  "logo_url": "https://api.apis.guru/v2/cache/logo/https_1forge.com_assets_images_f-blue.svg",
  "contact_name": "1Forge",
  "contact_email": "contact@1forge.com",
  "contact_url": "http://1forge.com",
  "license_name": null,
  "license_url": null,
  "terms_url": null,
  "external_docs_url": null,
  "homepage": "https://api.apis.guru/v2/specs/1forge.com/0.0.1.json",
  "is_unofficial": false,
  "added": "2017-05-30T08:34:14Z",
  "updated": "2017-06-27T16:49:57Z",
  "path_count": null,
  "operation_count": null,
  "source": "apis.guru",
  "scraped_at": "2026-07-28T18:06:02.869003Z"
}

path_count and operation_count are null unless countEndpoints is enabled. With it on, e.g. groundhog-day.com returns path_count: 5, operation_count: 5.

Provider mode items:

{ "type": "provider", "provider": "googleapis.com", "api_count": 281, "source": "apis.guru", "scraped_at": "..." }

All fields are nullable — the upstream directory is community-maintained and not every API ships a license, contact or category.

Pricing

Roughly $0.5 per 1,000 items on Apify compute. The directory is a single 8 MB JSON download, so a full 2,500-API run without enrichment finishes in well under a minute. Enabling countEndpoints adds one HTTP request per API — keep maxItems low when you use it.

Notes & limits

• Data comes from the public https://api.apis.guru/v2/ endpoints; APIs.guru is community-run and specs may lag behind the provider.
• Specs larger than 8 MB are skipped during enrichment (counts stay null).
• is_unofficial: true marks the ~25 specs APIs.guru maintains without the provider's blessing.
• The actor only reads public, non-personal metadata.

Related actors

• public-apis-scraper — the curated public-apis markdown list
• github-repos-scraper — repo metadata, stars and topics
• software-eol-scraper — end-of-life and support windows
• deps-dev-scraper — package dependency and security data

Local development

pip install -r requirements.txt
python -m src.main
python tests/test_smoke.py   # offline, stdlib only