Web & API AI Skills — HTTP, Scraping, REST, GraphQL, Auth (2026) | Skills and Servers
Subscribe
Category · 20 Skills

Web & API Skills — HTTP, Scraping, REST, GraphQL, Auth (2026)

Every AI Skill in our directory that touches HTTP, web pages, or APIs — request builders, scrapers, OpenAPI generators, GraphQL helpers, auth flow analyzers, rate-limit-aware crawlers. Each Skill has been tested against real endpoints (not toy examples) and documented for Claude, ChatGPT, Cursor, and Cline. Provider-agnostic.

20
Skills
6
Sub-types
5
Providers
100%
Human-tested
SS Curated by Skills and Servers Editorial Last updated August 14, 2026

1What this category covers

Web & API Skills accept a URL, an endpoint spec, a cURL command, an HTTP transcript, or a web page and produce something useful — a working request, a parsed response, an OpenAPI spec, a scraped dataset, a GraphQL schema, an auth flow diagram. Under the hood most of these Skills wrap standard HTTP libraries (requests, httpx, node-fetch) plus HTML parsers (beautifulsoup, cheerio, playwright).

The value the Skill layer adds is protocol correctness. Plain LLM output for HTTP work often gets details wrong — wrong Content-Type, missing charset, incorrect auth header formats, wrong OpenAPI structure. Skills enforce the details: they validate against schemas, follow RFC-required behavior, handle retries and rate limits properly. For anything that needs to work against real APIs, this discipline matters.

What lives here vs. elsewhere: if the Skill's output is HTTP-related or web-related, it belongs here. If it's application-level API endpoint generation (writing FastAPI or Express handlers), see Development & Code. If it's a research workflow that happens to use APIs, see Research & Analysis. For workflows chaining multiple API calls with logic between them, see Custom Workflows.

2When you should reach for a Web & API Skill

  • You want to test an API endpoint but don't have time to figure out the right auth headers and body format.
  • You have a cURL command from documentation and need it as code in your language.
  • You're integrating with an undocumented (or poorly-documented) API and need to reverse-engineer the shape.
  • You want to scrape a web page or crawl a site and need respectful (rate-limited, robots.txt-aware) scraping code.
  • You're building an OpenAPI or GraphQL spec from an existing API and need help generating a schema that matches actual behavior.

3Featured Skills this month

Our editorial picks — heaviest tested, highest reader ratings.

Editor's pick

cURL to Code Converter

The workhorse of the category. Paste any cURL command; get equivalent code in Python (requests/httpx), JS (fetch/axios), Rust (reqwest), Go, or shell. Preserves headers, auth, body format.

HTTP·Tested Aug 10·Sana K.
Most requested

OpenAPI Spec Generator from Endpoints

Point at a live API; get an OpenAPI 3.1 spec generated from actual response shapes. Uses real requests + JSON Schema inference. Iterative — improve the spec by feeding in more endpoints.

API Docs·Tested Aug 8·Ahmed R.
Highest reliability

Rate-Limit-Aware Web Crawler

Crawl a site respectfully — reads robots.txt, respects sitemap.xml, backs off on 429s, uses adaptive concurrency. Refuses obvious anti-scraping violations.

Scraping·Tested Aug 6·Sana K.

4All 20 Web & API Skills

Filter by sub-type, search by name, sort by recency or alphabetical.

Showing 20 of 20 Skills
HTTPAug 10

cURL to Code Converter

Paste cURL; get equivalent code in Python, JS, Rust, Go, shell. Preserves headers, auth, body format.

SKSana K.View →
HTTPAug 4

HAR File Analyzer

Analyze a HAR (HTTP Archive) file: waterfall, slowest requests, redirects, third-party breakdown, cookie flows.

ARAhmed R.View →
GraphQLAug 8

GraphQL Schema Explorer

Given a GraphQL endpoint, introspect the schema and explain it in plain English. Suggests useful queries.

ARAhmed R.View →
API DocsAug 6

OpenAPI to Markdown Docs

Generate readable Markdown API documentation from an OpenAPI spec. Includes request/response examples per endpoint.

SKSana K.View →
API DocsAug 4

SDK Generator from OpenAPI

Generate a typed client SDK (Python, TypeScript, Go) from an OpenAPI spec. Handles pagination, retries, auth.

ARAhmed R.View →

5Related MCP servers

MCP servers our team pairs with Web & API Skills for full workflows.

MCP Server

GitHub MCP Server

Pairs with OpenAPI and SDK Skills — commit generated specs and clients directly.

MCP Server

GitLab MCP Server

GitLab counterpart for teams on GitLab.com or self-hosted GitLab.

6Related Learn guides

Longer-form tutorials for the concepts behind these Skills.

7Comparisons worth reading

Head-to-head reviews that touch on Web & API workflows.

8Frequently asked questions

Read-only Skills (HTTP inspection, cURL conversion, HAR analysis, OpenAPI generation from responses) are generally safe against production — they only make requests you're already authorized to make. Write-related Skills (spec-vs-actual diff, endpoint tester) should be pointed at staging or dedicated test environments, not production, because they may make many test requests and hit rate limits. See per-Skill safety notes.
Auth credentials go in the input, not the Skill itself. For interactive workflows, paste the token or the full Authorization header. For automated workflows, store credentials in env vars or a secure store and reference them. Skills never save auth to disk or log them. The OAuth Flow Explainer can help you understand and set up OAuth for a new API before wiring auth into other Skills.
For most API work, Claude Opus 4.7 and GPT-5 are near-tied at the top. Claude is slightly better at reasoning across large OpenAPI specs and generating consistent SDKs; GPT-5 is slightly better at pattern-matching cURL commands to specific libraries. DeepSeek V4 is a great cost-conscious choice for high-volume mechanical work (cURL conversion, HAR analysis). See each Skill's Provider Compat table.
Some, but not aggressively. The Rate-Limit-Aware Web Crawler and JavaScript-Rendered Page Scraper Skills handle standard rate limiting and JS rendering. They do NOT try to bypass CAPTCHAs, cloak user agents, rotate residential proxies, or otherwise circumvent anti-bot defenses. If a site has strong anti-scraping, treat it as a signal that they don't want to be scraped — respect that. For legitimate use cases (your own site, or explicit permission), the Skills work well within respectful bounds.
Different roles. Postman/Insomnia are interactive API clients with GUIs, collections, environments, and team collaboration features. These Skills are code-generation and analysis tools that work inside your AI chat. They complement each other — use Postman for interactive API exploration and testing, use Skills for turning API concepts into code, generating specs, or debugging HTTP traces. The Postman Collection Generator Skill bridges the two by producing importable Postman collections from other inputs.

Working on a real Web & API problem?

Our sister sites cover errors and code recipes across the AI ecosystem. Bookmark them for when a Skill hits an edge case.

Visit AI Error Hub
SS
Curated by Skills and Servers Editorial
Editor-in-chief: Ahmed R. · Co-lead: Sana K.
Last updated August 14, 2026. Next full review of this category scheduled within 30 days. Every Skill entry is tested by a human on the platforms listed before publication.
Suggest a Skill

Share with

Twitter/XPinterestRedditMastodonThreadsEmailPrintFlipboardFlipboardHacker NewsLineWeChatSnapchatMixMixPocketGoogle MessagesSMSTumblrTumblrXingXINGVKVKontakteBufferBufferYummlyNextdoorGrocyGabFacebookWhatsAppLinkedInTelegramMessengerBluesky