RapidAPI MCP Skill

Version: 2.0.0
Last Updated: 2026-02-14
Author: OpenClaw
Description: Unified skill for accessing multiple RapidAPI services with modular architecture

---

Overview

This skill provides unified access to multiple RapidAPI-powered services. Each API is configured separately in the apis/ directory, allowing for easy addition, removal, and updates of APIs without modifying the core skill logic.

Features

• Modular Architecture: Each API has its own config file
• API Selector: Automatic routing to the correct API based on user intent
• Health Check: Test all APIs with a single command
• Subscription Detection: Automatic detection of subscription requirements
• API Discovery: Find and suggest APIs for unknown requests
• Smart Selection: Choose best API based on relevance, popularity, and free tier

---

Setup

Quick Start (Free APIs - Work Immediately!)

These APIs work without any API key:

# Create config directory
mkdir -p ~/.config/openclaw

# Test Open-Meteo (Weather)
curl "https://api.open-meteo.com/v1/forecast?latitude=52.5&longitude=13.4&current_weather=true"

# Test OpenLibrary (Books)
curl "https://openlibrary.org/search.json?q=python"

# Test PoetryDB
curl "https://poetrydb.org/title/love"

# Test Gutendex (Free ebooks)
curl "https://gutendex.com/books?search=alice"

# Test Bible
curl "https://bible-api.com/john+3:16"

Adding Your RapidAPI Key (Optional)

For premium APIs (YouTube, Twitter, Reddit, etc.), add your RapidAPI key:

1. Get a free key at https://rapidapi.com/auth/signup
2. Create config:

mkdir -p ~/.config/openclaw
echo '{"rapidapi_key": "YOUR_KEY_HERE"}' > ~/.config/openclaw/api-keys.json

Or set as environment variable:

export RAPIDAPI_KEY="YOUR_KEY_HERE"

# Run setup script to check all APIs
openclaw skill run rapidapi-mcp --setup

---

API Selector Prompt

When the user requests data from a specific service, use this prompt to route to the correct API:

"Which API do you need?"

🌟 FREE (work without key):

• Open-Meteo → openmeteo - Free global weather (best!)
• Open Library → openlibrary - Book search by title, author, ISBN
• Gutendex → gutendex - Project Gutenberg books
• PoetryDB → poetrydb - Poems and poetry
• Bible → bible - Holy Bible verses
• CoinGecko → coingecko - Free crypto prices

🔑 NEEDS RAPIDAPI KEY:

• YouTube → youtube - Videos, channels, trending
• YouTube (YT-API) → yt-api - Extended YouTube API with download, comments, hashtags, suggestions
• YouTube Captions → youtube-captions - Captions, transcripts, subtitles in multiple languages
• Twitter/X → twitter - Tweets, profiles, trends
• Reddit → reddit - Posts, comments, subreddits
• News → news - Latest headlines
• Weather → weather - Current weather, forecasts
• Streaming → streaming - Where to watch movies/shows
• Instagram → instagram - Profiles, posts, stories
• Wikipedia → wikipedia - Wikipedia articles
• OpenStreetMap → openstreetmap - Geocoding, maps
• Wikidata → wikidata - Structured facts
• Archive → archive - Internet Archive
• Harry Potter → harrypotter - HP characters, spells
• Nobel Prize → nobel - Nobel laureates

---

Available Tools

1. rapidapi_call

Call any RapidAPI endpoint with automatic routing.

Parameters:

• api (string): API name from config (youtube, twitter, reddit, news, weather, streaming, crypto, instagram)
• endpoint (string): Endpoint name as defined in API config
• params (object): Query parameters for the endpoint

Example:

{
  "api": "youtube",
  "endpoint": "search",
  "params": {
    "query": "openai",
    "maxResults": 5
  }
}

2. rapidapi_check

Check health/subscription status of all configured APIs.

Parameters:

• api (string, optional): Specific API to check, or "all" for all APIs

Example:

{
  "api": "all"
}

3. rapidapi_setup

Interactive setup to configure API subscriptions.

Parameters:

• action (string): "check", "configure", or "status"

Example:

{
  "action": "check"
}

4. rapidapi_discover

Find APIs for unknown requests. Searches RapidAPI for relevant APIs when no known API is configured.

Parameters:

• query (string): What the user wants to do (e.g., "weather", "reddit", "news")
• max_results (number): Maximum number of APIs to return (default: 5)

Example:

{
  "query": "weather forecast",
  "max_results": 5
}

Selection Logic:

1. Check if API is already configured → use it
2. If not → search RapidAPI
3. Rank results by: Relevance → Free Tier → Popularity
4. Return top 3 with subscription links

---

Token Optimization for Browser Scraping

When scraping RapidAPI documentation pages via browser, use these optimizations to reduce token usage:

1. Use compact: true for Snapshots

{
  "action": "snapshot",
  "compact": true,
  "profile": "chrome"
}

Result: Reduces ~1.5k tokens → ~200 tokens per snapshot

2. Use Screenshot for Visual Parameters

When you only need to see parameter values visually:

{
  "action": "screenshot",
  "profile": "chrome"
}

3. Selective Element Access

Instead of full page snapshots, use targeted selectors when possible.

---

Available APIs

IMDb146

• Host: imdb146.p.rapidapi.com
• Description: Movie database - titles, names, videos, news
• Documentation: https://rapidapi.com/Glavier/api/imdb146

Endpoint	Method	URL	Parameters
Find	GET	/v1/find/	query (string, required)
Title Details	GET	/v1/title/	id (string, required) - Title ID
Name Details	GET	/v1/name/	id (string, required) - Name ID
Video Details	GET	/v1/video/	id (string, required) - Video ID
News Details	GET	/v1/news/	id (string, required) - News ID

YouTube (YT-API)

• Host: yt-api.p.rapidapi.com
• Description: YouTube Data + Download API - videos, channels, search, trending, playlists, comments, download URLs
• Documentation: https://rapidapi.com/ytjar/api/yt-api
• Pricing: Basic (Free), Pro ($51/mo), Ultra ($144/mo), Mega ($240/mo)

Endpoint	Method	Path	Parameters
Video Info	GET	/video	id (required), pretty, geo, lang
Related Videos	GET	/related	id (required), token, pretty, geo, lang
Trending	GET	/trending	geo, type (now/music/games/movies), pretty, lang
Search	GET	/search	query (required), type, sort, duration, upload_date, features, token, pretty, geo, lang
Playlist	GET	/playlist	id (required), token, pretty, geo, lang
Channel	GET	/channel	id (required), sort_by (newest/oldest/popular), token, pretty, geo, lang
Comments	GET	/comments	id (required), token, pretty, geo, lang
Hashtag	GET	/hashtag	id (required), pretty, geo, lang
Resolve URL	GET	/resolve	url (required), pretty
Download/Stream	GET	/dl	id (required), cgeo, pretty (quota: 6 units)
Home Feed	GET	/home	pretty, geo, lang
Suggestions	GET	/suggest	query (required), pretty
Hype	GET	/hype	pretty, geo, lang

YouTube Captions

• Host: youtube-captions-transcript-subtitles-video-combiner.p.rapidapi.com
• Description: YouTube Captions, Transcripts, Subtitles - Download in SRT, VTT, XML, JSON; translate to any language
• Documentation: https://rapidapi.com/nikzeferis/api/youtube-captions-transcript-subtitles-video-combiner
• Pricing: Basic (Free), Pro ($3.87/mo), Ultra ($13.87/mo), Mega ($34.98/mo)

Endpoint	Method	Path	Parameters
Video Info	GET	/get-video-info/{videoId}	videoId (required), format, response_mode
Language List	GET	/language-list/{videoId}	videoId (required), format, response_mode
Download SRT	GET	/download-srt/{videoId}	videoId, language (required), response_mode
Download XML	GET	/download-xml/{videoId}	videoId, language (required), response_mode
Download JSON	GET	/download-json/{videoId}	videoId, language (required), response_mode
Download WebVTT	GET	/download-webvtt/{videoId}	videoId, language (required), response_mode
Download All	GET	/download-all/{videoId}	videoId, format_subtitle, format_answer, response_mode

YouTube

• Host: youtube-v31.p.rapidapi.com
• Endpoints: search, videoDetails, channelDetails, trending
• Use Case: Video search, channel info, trending content

Twitter/X

• Host: twitter135.p.rapidapi.com
• Endpoints: userTweets, userByUsername, search, trends
• Use Case: Tweet retrieval, user profiles, trending topics

Reddit

• Host: reddit23.p.rapidapi.com
• Endpoints: subredditPosts, postComments, subredditSearch, trendingSubreddits
• Use Case: Reddit content, subreddit browsing, comments

News

• Host: newsdata.io
• Endpoints: latestNews, newsArchive, sources
• Use Case: Headlines, article search, news sources

Weather

• Host: weatherapi-com.p.rapidapi.com
• Endpoints: currentWeather, forecast, history, astronomy
• Use Case: Weather data, forecasts, astronomy info

Streaming

• Host: streaming-availability.p.rapidapi.com
• Endpoints: search, getDetails, getCountries, getLanguages
• Use Case: Find where to stream movies/shows

Crypto

• Host: coinranking1.p.rapidapi.com
• Endpoints: coins, coinDetails, coinHistory, exchanges, trending
• Use Case: Cryptocurrency prices, market data

Instagram

• Host: instagram120.p.rapidapi.com
• Endpoints: userProfile, userPosts, hashtagSearch, mediaDetails
• Use Case: Instagram profiles and media retrieval

Books

• API: Google Books (free, no key needed)
• Endpoints: search, by_isbn
• Use Case: German books, by genre, date, ISBN

Wikipedia

• API: de.wikipedia.org (free, no key needed)
• Endpoints: search, page_content, page_details
• Use Case: German Wikipedia articles, search

OpenStreetMap

• API: nominatim.openstreetmap.org (free, no key)
• Endpoints: search, reverse, lookup
• Use Case: Geocoding, location search

Open Library

• API: openlibrary.org (free, no key)
• Endpoints: search, book_by_isbn, author, cover
• Use Case: Book search, ISBN lookup, covers

Wikidata

• API: wikidata.org (free, no key needed)
• Endpoints: search, get_entity, sparql
• Use Case: Structured facts, entities

Archive

• API: archive.org (free, no key needed)
• Endpoints: search, metadata
• Use Case: Historical books, media, web archives

CoinGecko

• API: api.coingecko.com (free, no key needed)
• Endpoints: coins, coin_detail, trending, global, simple_price, search
• Use Case: Cryptocurrency prices, market data, trending

Open-Meteo

• API: api.open-meteo.com (free, no key needed)
• Endpoints: forecast
• Use Case: Global weather, forecasts

Gutendex

• API: gutendex.com (free, no key needed)
• Endpoints: books, book_by_id
• Use Case: Project Gutenberg free books

PoetryDB

• API: poetrydb.org (free, no key needed)
• Endpoints: random, author, title, search
• Use Case: Poems, poets

Harry Potter API

• API: potterapi-fedeperin.vercel.app (free, no key needed)
• Endpoints: books, characters, spells, houses
• Use Case: Harry Potter data

Bible

• API: bible-api.com (free, no key needed)
• Endpoints: verse, random
• Use Case: Bible verses

Nobel Prize

• API: api.nobelprize.org (free, no key needed)
• Endpoints: prizes, laureates
• Use Case: Nobel Prize data

---

API Configuration Files

All API configurations are stored in /opt/homebrew/lib/node_modules/openclaw/skills/rapidapi-mcp/apis/:

apis/
├── youtube.json
├── twitter.json
├── reddit.json
├── news.json
├── weather.json
├── streaming.json
├── crypto.json
└── instagram.json

Each config file contains:

• name: Display name
• host: RapidAPI host
• description: What the API does
• rapidapi_key: API key for authentication
• endpoints: Available endpoints with paths and params
• subscription_url_pattern: URL pattern for subscription

API Selection Criteria

When recommending APIs, use this priority order:

1. Relevanz - Does the API do what the user needs?
2. Free Tier - How many calls are free/month?
3. Popularität - How many users? (proxy for reliability)
4. Rate Limits - Enough for normal use case?

API Discovery Flow

User Request → Check known APIs → Not found? 
→ Search RapidAPI → Rank by criteria → User chooses
→ Subscribe → Ready to use

---

Example Usage

Key Loading Order

When making API calls, the skill loads the key in this order:

1. ~/.config/openclaw/api-keys.json → rapidapi_key
2. Environment variable RAPIDAPI_KEY
3. Config file value (should be null)

Search YouTube for "AI tutorials"

api: rapidapi_call
params: {
  "api": "youtube",
  "endpoint": "search",
  "params": {
    "query": "AI tutorials",
    "maxResults": 10
  }
}

Get current weather for Berlin

api: rapidapi_call
params: {
  "api": "weather",
  "endpoint": "currentWeather",
  "params": {
    "q": "Berlin"
  }
}

Get trending crypto

api: rapidapi_call
params: {
  "api": "crypto",
  "endpoint": "trending",
  "params": {}
}

Check all API subscriptions

api: rapidapi_check
params: {
  "api": "all"
}

---

Subscription Management

How Subscriptions Work

1. Most RapidAPI endpoints require a subscription
2. Free tier: Limited requests/month
3. Paid tier: Higher limits, more endpoints
4. Each API has a subscription_url_pattern pointing to its RapidAPI page

Check Subscription Status

# Check all APIs
openclaw skill run rapidapi-mcp --check

# Check specific API
openclaw skill run rapidapi-mcp --check youtube

Subscribe to an API

1. Run the check command to see which APIs need subscriptions
2. Visit the subscription URL from the API config
3. Select a plan (Free or Paid)
4. Update the API key in the config if needed

---

Error Handling

Common Errors

Error	Cause	Solution
401 Unauthorized	Invalid API key	Update key in config
403 Forbidden	Subscription required	Subscribe on RapidAPI
429 Too Many Requests	Rate limit exceeded	Wait or upgrade plan
404 Not Found	Invalid endpoint	Check endpoint name

Best Practices

1. Check API health before making calls
2. Handle rate limits gracefully
3. Cache responses when possible
4. Use specific endpoints instead of generic ones

---

File Structure

rapidapi-mcp/
├── SKILL.md              # This file
├── README.md             # Original documentation
├── setup.sh              # Setup and check script
└── apis/
    ├── youtube.json
    ├── twitter.json
    ├── reddit.json
    ├── news.json
    ├── weather.json
    ├── streaming.json
    ├── crypto.json
    └── instagram.json

---

Notes

• This skill uses the RapidAPI key from: ~/.config/openclaw/api-keys.json or environment variable RAPIDAPI_KEY
• Some APIs may require additional RapidAPI subscriptions
• Free tier limits apply to most APIs
• Check individual API docs for rate limits and capabilities