ScrapeCreators

Why this is showing up on feeds: Riley Brown / Agent Native marketing masterclass calls out a “ScrapeCreators skill” as the first daily tool for agents that run marketing — pull social data for outreach, ideation, niche research, ads, UGC, and X research.

• @rileybrown on ScrapeCreators
• Product: scrapecreators.com · docs: docs.scrapecreators.com
• Founder: @adrian_horning_ (indie / “run by a human”)

This page is an independent field note for our stack — not an affiliate pitch.

What it is (in plain language)

ScrapeCreators is a hosted social media data API. You send a simple HTTP request with an API key; you get structured JSON about public profiles, posts, videos, comments, transcripts, search results, and several ad libraries — without maintaining Playwright farms, rotating proxies, or platform OAuth apps yourself.

Think of it as the social counterpart to SEO data APIs like DataForSEO: pay for structured data so agents and apps can research and act, not babysit scrapers.

Why agents (and marketers) care

In the Agent Native “9 skills” marketing masterclass, skill #1 is scrape social; #2 is turn a creator into a reusable skill; #3 is scrape competitor ads. ScrapeCreators is the plumbing for those moves.

Platform coverage (high level)

One API surface across many networks (counts move; check docs). Typical clusters:

• Short video: TikTok (incl. shop / ads library), Instagram reels/posts, YouTube / Shorts, Snapchat, Kwai…
• Social graphs & posts: X/Twitter, LinkedIn, Facebook, Threads, Bluesky, Reddit, Truth Social…
• Ad libraries: Facebook/Meta, TikTok, Google, LinkedIn
• Audio / other: Spotify, Apple Music, SoundCloud, Twitch, GitHub, Pinterest, link-in-bio pages…

Common resource types: profile, posts/videos, comments, transcripts, search, trending, ads.

How pricing works (conceptually)

Model (per their site; confirm current numbers before buying):

• Credits — buy packs; claims of credits never expire
• Free starter — small credit allotment to try (e.g. 100 free credits)
• Cache hits free — if they already have the result cached, 0 credits
• No rate limits (as marketed) — scale concurrent requests; you still manage cost
• All APIs on plans — not “pay extra per platform” silos in their pricing story

Practical habit: cache yourself too (don’t re-pull the same profile every minute) and prefer “research batch” jobs over chatty agent loops.

How to use it — four ways in

1. Raw HTTP (simplest mental model)

Base: https://api.scrapecreators.com · header: x-api-key: YOUR_API_KEY

curl "https://api.scrapecreators.com/v1/tiktok/profile?handle=SOME_HANDLE" \
  -H "x-api-key: YOUR_API_KEY"

curl "https://api.scrapecreators.com/v1/instagram/profile?handle=SOME_HANDLE" \
  -H "x-api-key: YOUR_API_KEY"

Sign up → get key in the dashboard → hit docs playground for the exact path you need.

2. MCP (Cursor / Claude Desktop / MCP clients)

Official MCP package so agents get endpoints as tools. Config shape (from their docs; verify package name if it changes):

{
  "mcpServers": {
    "scrape-creators": {
      "command": "npx",
      "args": ["@scrape-creators/mcp"]
    }
  }
}

Docs: integrations/mcp — set API key via env as documented.

3. Agent skill (Codex / Claude Code / Cursor)

This is what the marketing video means by the “ScrapeCreators skill”: a first-party skill that teaches the agent which endpoints exist, parameters, and response shapes so it stops inventing fake scrapers.

• Skill docs: agent-skill
• Machine-readable docs for agents: llms.txt, llms-full.txt, openapi.json

In practice: install skill → put key in secrets → prompt: “Pull last 20 reels from @…, rank by views, summarize hooks.”

4. CLI

Terminal / cron / CI friendly (@scrapecreators/cli on npm per their site). Example pattern: scrape-creators tiktok profile @handle — good for batch scripts and HEOS-style automation.

Also: n8n community node and Apify actors if you live in those tools.

Suggested first workflows (do these this week)

1. Hello world: one TikTok or Instagram profile JSON → save to a file; confirm credits moved.
2. Competitor ad pull: Meta or TikTok ad library search for a keyword / company → 10 creatives → agent summary of offers + hooks.
3. Creator shortlist: search users/hashtag in a niche → table: handle, followers, last 3 post topics, contact guess from bio/link-in-bio.
4. Transcript → content: one video transcript endpoint → outline for your own post or comic episode.
5. Agent skill daily: “What did these 5 competitors post this week?” as a scheduled agent job.

ScrapeCreators vs neighbors

Their own framing: Firecrawl = any site HTML; ScrapeCreators = social-specialized structured data. Use both in a serious marketing agent stack (skill #5 in that masterclass is Firecrawl for brand assets).

How it fits our stack

Compliance & hygiene (don’t skip)

• Public data only — respect platform ToS and local law for your use case (outreach, storage, resale).
• Don’t spam — enrichment ≠ permission to blast DMs.
• Store carefully — PII retention policies; purge lists you don’t need.
• Attribute honestly — competitive research for learning ≠ copying ads/assets wholesale.
• Budget agents — cap credits per job so a runaway loop doesn’t burn a pack.

Quick start checklist

• □ Account at app.scrapecreators.com
• □ API key in a secrets manager (not committed to git)
• □ One successful curl profile request
• □ Install MCP or agent skill for daily driver agent
• □ Define 1 research job (e.g. “5 competitors’ ads this week”)
• □ Folder/table for outputs (CSV / HEOS vault / Notion)
• □ Credit budget note in the job brief

Related

Living notes

Pricing, package names, and endpoint versions change — always confirm against official docs before production. Next edit: paste our preferred “daily research” agent prompt once we run it for real.