46 endpoints · 7 platforms · one key

Scrape any social
platform with one API

Transcripts, posts, profiles, comments and stats from YouTube, TikTok, Instagram, Facebook, X and LinkedIn. Send a URL, get clean JSON back. No headless browsers, no proxy pool, no scrapers to keep alive.

Start freeBrowse the endpoints

25 free credits · no card required

POST /v1/youtube/transcript

200 OK

{
  "platform": "youtube",
  "video_id": "dQw4w9WgXcQ",
  "title": "How we cut inference cost by 60%",
  "duration_seconds": 734,
  "language": "en",
  "transcript": [
    { "start": 0.0,  "text": "So the bill arrived and" },
    { "start": 3.4,  "text": "nobody wanted to open it." }
  ],
  "summary": "A walkthrough of batching, caching",
  "key_points": ["Batch small calls", "Cache prompts"],
  "credits_used": 1
}
1 credit used1.9s
  • YouTube
  • TikTok
  • Instagram
  • Facebook
  • X / Twitter
  • LinkedIn
  • Video & audio files

The catalogue

46 endpoints. Six platforms and anything with a media URL.

Each one takes a public link and returns a documented JSON shape. Where a concept exists on more than one platform — comments, profile details, search — the response fields match, so the code you write for YouTube already works for TikTok.

YouTube

11

Long-form and Shorts, including full transcripts.

  • transcript
  • summary
  • video-details
  • comments
  • search
  • channel-details
  • channel-videos
  • playlist-videos
  • shorts-transcript
  • shorts-details
  • download

TikTok

9

Videos, creators and hashtag discovery.

  • transcript
  • summary
  • video-details
  • comments
  • search
  • hashtag-search
  • channel-details
  • channel-videos
  • download

Instagram

9

Reels and profile grids with engagement counts.

  • reel-transcript
  • reel-summary
  • reel-details
  • comments
  • reels-search
  • profile-details
  • profile-posts
  • profile-reels
  • download

Facebook

5

Public video posts and Pages.

  • transcript
  • summary
  • post-details
  • comments
  • page-details

X / Twitter

5

Tweets, threads and profile timelines.

  • tweet
  • thread
  • profile
  • profile-tweets
  • video-transcript

LinkedIn

5

People, companies and post engagement.

  • profile
  • company
  • company-posts
  • post-details
  • video-transcript

Video & audio files

2

Any public media URL, no platform required.

  • transcript
  • summary

Something missing?

We add platforms and endpoints based on what people actually ask for. Tell us what you are trying to pull and we will tell you whether it is on the roadmap.

Request an endpoint

How it works

Four steps, and none of them involve a browser.

  1. 01

    Pick an endpoint

    One route per job — transcript, comments, profile, search. No graph to learn.

  2. 02

    Send the URL

    Paste the public link exactly as you copied it. We resolve short links, Shorts, Reels and mobile URLs.

  3. 03

    Get JSON back

    A stable, documented shape per endpoint. Same fields on every platform where the concept exists.

  4. 04

    Ship it

    Feed it to an LLM, a dashboard, a Zap or a cron job. One credit per successful call.

Why an API instead of a scraper

The hard part was never the first request.

It was the Tuesday six weeks later when the markup changed, the proxy pool got flagged and the job that fed your dashboard started returning empty arrays without telling anyone.

No scrapers to maintain
Layouts change, tokens rotate, endpoints move. That becomes our on-call rotation instead of yours.
Consistent JSON
Comments look the same on TikTok as on YouTube. Write your parsing once and point it at any platform.
Transcripts everywhere
Native captions when they exist, speech-to-text when they don't — with timestamps either way.
AI-ready output
Summaries, key points and topic tags are first-class responses, not something you post-process.
One bearer token
No OAuth dance, no per-platform app review, no rotating cookies. A header and a URL.
Honest failures
Typed error codes and no charge for a failed call. Deleted, private and geo-blocked are distinct answers.

Wiring it up

A header and a URL. Or no code at all.

Call it with curl, hand it to an agent over MCP, or drop it into a Zap. Same endpoints, same credits, same responses — pick whichever front door suits the job.

  • Zapier

    Native app

  • Make

    Native modules

  • n8n

    Community node

  • MCP server

    Claude, Cursor

  • Agent Skills

    Drop-in skill pack

  • REST

    curl, any language

Shell
curl https://api.scrapesocial.io/v1/tiktok/comments \
  -H "Authorization: Bearer $SCRAPESOCIAL_KEY" \
  -G --data-urlencode "url=https://tiktok.com/@user/video/123"
MCP client config
{
  "mcpServers": {
    "scrapesocial": {
      "url": "https://mcp.scrapesocial.io",
      "headers": {
        "Authorization": "Bearer $SCRAPESOCIAL_KEY"
      }
    }
  }
}

Who uses it

Built for anything that needs to read the internet.

AI apps & agents
Give a model the actual contents of a link instead of a URL it can't open.
Social listening
Track mentions, comments and sentiment across platforms on one schedule.
Influencer research
Pull follower counts, posting cadence and real engagement before you sign a deal.
Content repurposing
Turn one long video into threads, newsletters, clips and show notes.
Market & trend research
Sweep hashtags and search results on a cron and watch what moves.
Analytics dashboards
Back your reporting with post-level data rather than screenshots.
No-code automations
Wire the same endpoints into Zapier, Make or n8n without writing a line.
Internal tooling
Stop maintaining the one Python scraper nobody on the team wants to own.

Pricing

Pay for calls that work.

One credit per successful request, across all 46 endpoints. Failed calls cost nothing. Change plan whenever the shape of your usage changes.

Free

Enough to run every endpoint you care about against your own URLs.

$0

40 req/min

  • 25 credits, once
  • Every endpoint, every platform
  • No card required
Create an account

Starter

For a side project or a single production workflow.

$29/month

120 req/min

  • 12,000 credits per month
  • Every endpoint, every platform
  • Zapier, Make and n8n steps
  • MCP server access
  • Email support

Growth

Popular

For products with real traffic and scheduled jobs.

$59/month

240 req/min

  • 30,000 credits per month
  • Every endpoint, every platform
  • Zapier, Make and n8n steps
  • MCP server access
  • Webhook delivery for long jobs
  • Priority support

Scale

For platforms pulling social data continuously.

$149/month

480 req/min

  • 90,000 credits per month
  • Every endpoint, every platform
  • Zapier, Make and n8n steps
  • MCP server access
  • Webhook delivery for long jobs
  • Higher concurrency ceiling
  • Priority support with an SLA

Need a higher ceiling, a private deployment or an invoice? Talk to us about a custom plan.

Questions

Before you sign up.

What exactly is ScrapeSocial?

A single REST API for reading public social media data. You send a post, profile or video URL to an endpoint and get structured JSON back — transcripts, comments, engagement counts, profile details and AI summaries — across YouTube, TikTok, Instagram, Facebook, X and LinkedIn.

Do I need to build or host anything?

No. There are no scrapers, browsers, proxies or queues to run. Everything is an HTTPS request with a bearer token, so it works from a Lambda, a laptop or a no-code tool equally well.

How does credit-based pricing work?

Every plan comes with a monthly credit allowance. A successful API call spends one credit; a few of the heavier endpoints, such as speech-to-text transcription of a long video, spend more, and that cost is documented per endpoint. Failed calls are never charged.

What happens if I run out of credits?

Calls start returning a quota error rather than silently failing or auto-billing you. You can upgrade at any time and the new allowance applies immediately; unused credits do not roll over between months.

Can I try it before paying?

Yes. Every account starts with free credits and no card on file, which is enough to run each endpoint you care about against your own URLs and see the exact response shape.

Is this only for developers?

The API is the product, but you don't have to write code to use it. Native Zapier, Make and n8n steps cover the common workflows, and an MCP server lets agents like Claude Code call the endpoints directly.

What data can it access?

Public data only. Anything that requires logging in as a particular user — private accounts, DMs, follower lists behind a login, unlisted posts — is out of scope, and the API tells you so explicitly instead of guessing.

How fast is it?

Metadata and comment endpoints typically return in one to three seconds. Transcription of a long video depends on its length; those endpoints stream progress so you are not left holding an open socket.

Delete the scraper.

46 endpoints, one bearer token and 25 free credits to prove it does what you need before you pay anything.

Start freeSee pricing