A social media API toolkit for AI agents. These skills let Claude, ChatGPT, Cursor, or any custom agent loop read public data from 33 platforms — the TikTok API for video transcripts, the Instagram API for reels and hashtags, the LinkedIn API for profiles and company data, the YouTube API for transcripts and comments, plus Reddit, X, Facebook, Amazon, Google and more — through one key and one JSON envelope. There is also a hosted MCP server if you would rather install nothing at all.
227 endpoints across 33 platforms
Free to start: signing up gives you 100 credits, no card required. Plain Node.js. No dependencies. MIT-0 licensed.
Most people want the catch-all skill — one skill that reaches all 33 platforms:
# OpenClaw (via ClawHub)
npx clawhub@latest install social-media-api
# Hermes Agent
hermes skills install skills-sh/nikhonit/social-media-api-skills/skills/social-media-api
# Generic agent skills (Claude Code, Cursor, Cline) — installs every skill in the repo
npx skills add nikhonit/social-media-api-skillsOr install one platform, when you already know which you need:
npx clawhub@latest install tiktok-api
npx clawhub@latest install instagram-api
npx clawhub@latest install linkedin-api
npx clawhub@latest install youtube-apiAny skill name from the table below works. The per-platform skills carry argument validation, per-endpoint costs and guidance on when not to spend credits; the catch-all trades that for reach.
| Skill | Platform | Endpoints | What it covers |
|---|---|---|---|
| amazon-api | Amazon | 4 | Product detail, product search, reviews and seller listings. |
| app-store-api | App Store | 6 | App metadata, reviews, keyword search and category charts. |
| bluesky-api | Bluesky | 3 | Public profiles, post timelines and keyword search. |
| facebook-api | 23 | Transcripts, page stats, comments, reels, ads, marketplace and events. | |
| github-api | GitHub | 11 | Repos, commits, contributors, issues, releases, profiles and trending. |
| google-api | 5 | Search results, Business Profile info, reviews and company ads. | |
| google-finance-api | Google Finance | 3 | Quotes, historical price history and instrument search. |
| google-maps-api | Google Maps | 4 | Place details, place search, photos and contact discovery. |
| google-news-api | Google News | 1 | Keyword search across news headlines and sources. |
| google-play-api | Google Play | 4 | Android app metadata, reviews, keyword search and categories. |
| google-shopping-api | Google Shopping | 1 | Product search across merchants, with pricing. |
| google-trends-api | Google Trends | 2 | Interest over time and rising related queries. |
| hacker-news-api | Hacker News | 4 | Stories, comment threads, user profiles and keyword search. |
| instagram-api | 23 | Reel transcripts, profile and post stats, comments, hashtags and locations. | |
| kwai-api | Kwai | 3 | Creator profiles, post listings and single video lookups. |
| linkedin-api | 38 | Profiles and 20 sub-resources, companies, job and people search, posts and ads. | |
| linkme-api | Linkme | 1 | Resolve a Linkme page to its links and profile details. |
| linktree-api | Linktree | 1 | Resolve a Linktree page to its links and profile details. |
| naver-api | Naver | 4 | Blog, cafe, local and shopping search across Korea's main portal. |
| pinterest-api | 3 | Pin details, keyword search and per-URL save counts. | |
| polymarket-api | Polymarket | 1 | Prediction market listings and current odds. |
| reddit-api | 8 | Posts, comment threads, subreddit listings, user profiles and search. | |
| snapchat-api | Snapchat | 1 | Public creator profile lookups. |
| spotify-api | Spotify | 6 | Tracks, albums, artists, top tracks, playlists and search. |
| threads-api | Threads | 5 | Profiles, post timelines, post stats and keyword search. |
| tiktok-api | TikTok | 16 | Transcripts, summaries, channel stats, comments, hashtag and song search. |
| tiktok-shop-api | TikTok Shop | 5 | Product detail, reviews, shop catalogues and creator showcases. |
| tripadvisor-api | Tripadvisor | 2 | Hotel, restaurant and attraction search with review listings. |
| trustpilot-api | Trustpilot | 1 | Company review listings. |
| truth-social-api | Truth Social | 2 | Public profiles and user post timelines. |
| twitch-api | Twitch | 3 | Streamer profiles, clips and video listings. |
| twitter-api | X (Twitter) | 7 | Transcripts, AI summaries, post stats, profiles, tweets and search. |
| youtube-api | YouTube | 26 | Transcripts, captions, video and channel stats, comments, Shorts and search. |
| social-media-api | All platforms | 227 | Discover and call any endpoint across every platform. |
# 1. Get a key — 100 free credits, no card
open https://scrapersocial.com/signup
# 2. Export it
export SCRAPERSOCIAL_KEY=sk_live_...
# 3. Run any script
node skills/github-api/scripts/get_profile.js torvaldsNote: no captured sample output is committed yet. The response shape shown below is taken from the OpenAPI spec — it is not a recorded call. Run
SCRAPERSOCIAL_KEY=... node tools/capture.mjsto record real output and regenerate.
Every response uses the same envelope:
{
"data": {},
"meta": { "count": 20, "next_cursor": "..." },
"request_id": "a18e4ef88ad4b00b"
}data is an object for single-entity endpoints and an array for paginated ones. meta appears on paginated responses only. Keep request_id — it is how support traces a call.
Errors follow the same discipline:
{
"error": {
"code": "invalid_url",
"message": "...",
"hint": "...",
"request_id": "..."
}
}Failed calls are not charged.
These are plain Node scripts with a documented contract, so anything that can run a command can use them.
- Claude Code / Claude Skills — install with
npx skills add nikhonit/social-media-api-skills - ChatGPT — expose a script as a function or run it in a code interpreter step
- Cursor, Cline, Windsurf — drop the repo in and let the agent read
SKILL.md - Any agent loop — call the scripts as subprocesses and parse stdout
If you would rather not vendor anything, ScraperSocial runs a hosted MCP server that reaches the same endpoints:
claude mcp add --transport http scrapersocial https://mcp.scrapersocial.com{
"mcpServers": {
"scrapersocial": {
"url": "https://mcp.scrapersocial.com",
"headers": { "Authorization": "Bearer sk_live_..." }
}
}
}It exposes 14 tools — twelve capability tools (get_transcript, get_summary, get_post_stats, get_comments, get_comment_replies, get_profile, get_channel_posts, search, search_hashtag, get_reviews, download_media, get_trending) plus list_endpoints and call_endpoint, which between them reach the entire surface. Auth is OAuth 2.1 PKCE or a bearer key.
Use the skills in this repo when you want per-platform guidance, argument validation, and control over exactly which endpoint runs. Use the MCP server when you want nothing installed.
- Clean JSON, one envelope for every platform
request_idon every response, including errorsx-cache: hit|missheader telling you whether it came from cache- Projection and format control —
fieldstrims the response,formatswitches betweenjson,csvandndjson - OpenAPI 3.1: scrapersocial.com/openapi.json
- Agent-readable summary: llms.txt · llms-full.txt
Every endpoint is GET https://api.scrapersocial.com/v1/{platform}/{capability}. Optional parameters on every call:
| Param | Effect |
|---|---|
limit |
Items on paginated endpoints. Default 20. Scripts clamp this to 50 — above that the API returns an async job instead of data. |
cursor |
next_cursor from a previous response's meta. |
fields |
Projection, e.g. fields=id,author.name. |
format |
json (default), csv, ndjson. |
fresh |
true bypasses the cache at the same credit cost. |
Every script reads one environment variable:
export SCRAPERSOCIAL_KEY=sk_live_...Manage keys at scrapersocial.com/app/keys. The scripts never print your key — error output carries the request_id instead. Keep it out of source control.
| Plan | Price | Credits | Rate limit |
|---|---|---|---|
| Free | $0 | 100 one-time at signup, no card | 20/min |
| Monthly | $5/mo | 1,000/month, top-ups $4/1k | 200/min |
| Annual | $54/yr | 12,000 upfront, top-ups $3/1k | 300/min |
| Enterprise | custom | custom | 1,500/min |
Credits are charged per item returned on paginated endpoints and per call otherwise. Costs range from 1 credit (GitHub, Hacker News, Bluesky, Linktree) to 42 (Instagram AI summary). Each skill's SKILL.md lists the exact cost of every endpoint it wraps. Cache hits cost the same as fresh calls and return in milliseconds. Failed calls are not charged.
Install tiktok-api, export SCRAPERSOCIAL_KEY, and run node skills/tiktok-api/scripts/get_transcript.js <video-url>. It returns the spoken content as JSON for 8 credits. Channel stats, comments, hashtag search and song lookups are separate scripts in the same skill.
This one starts free: 100 credits at signup with no card. That covers roughly 100 GitHub or Hacker News calls, or a dozen TikTok transcripts. After that, plans start at $5/month for 1,000 credits.
Two ways. Install the linkedin-api skill from this repo, or add the hosted MCP server with claude mcp add --transport http scrapersocial https://mcp.scrapersocial.com. Both cover profiles, companies, job and people search, posts and the ad library.
An MCP server exposes tools an AI agent can call directly, without you writing glue code. ScraperSocial's is at https://mcp.scrapersocial.com and exposes 14 tools covering all 33 platforms. Add it once and the agent can read social data on its own.
Yes — node skills/instagram-api/scripts/get_transcript.js <reel-url>, 40 credits. An AI summary of the same reel costs 42. TikTok transcripts are 8 credits and YouTube transcripts are 3, so check which platform you actually need.
Between 1 and 42 credits per call depending on the endpoint, with credits from $3–5 per thousand. GitHub, Hacker News, Bluesky and Linktree are 1 credit. Video transcripts and AI summaries are the expensive end. Every SKILL.md lists exact per-endpoint costs.
No. ScraperSocial is an independent product, not affiliated with any platform named here, and returns publicly available data only. Where a platform offers its own API — GitHub, Spotify, Reddit, Bluesky — that is often free and richer; these skills are most useful for the platforms that do not, and for using one key across all of them.
They need Node 20 or newer, because they use built-in fetch. There are no other dependencies and no build step. If you would rather not run Node at all, use the hosted MCP server.
Bug reports and pull requests are welcome at github.com/nikhonit/social-media-api-skills.
Note that skills/ and manifest.json are generated — edit tools/content/*.md and run node tools/generate.mjs. See CONTRIBUTING.md.
For API questions or a key problem, email support@scrapersocial.com with the request_id from the response.
MIT No Attribution (MIT-0). Use these however you like, no attribution required.
ScraperSocial is an independent product and is not affiliated with, endorsed by, or sponsored by any platform named in this repository. All product names, logos and brands are property of their respective owners. These skills return only publicly available data.