SocialCrawl MCP
The SocialCrawl MCP server gives an agent profiles, posts, reviews and prices from social networks and shops, with every call priced in credits
- 28 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The API key spends credits on every paid call, and for metered endpoints the price is known in advance only as a range
- `socialcrawl_manage` starts monitors, browser sessions and background jobs that keep spending credits
- Collected profiles and comments contain people's personal data, and platform rules limit collecting them
Install
- Works in
Manual install
claude mcp add --scope user socialcrawl -e SOCIALCRAWL_API_KEY=sc_your_key_here -- npx -y socialcrawl-mcpLocal run through npx. Replace sc_your_key_here with your key.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The API key spends credits on every paid call, and for metered endpoints the price is known in advance only as a range
- `socialcrawl_manage` starts monitors, browser sessions and background jobs that keep spending credits
- Collected profiles and comments contain people's personal data, and platform rules limit collecting them
This is third-party code. Review the repository files before installing.
What it does
The server connects an agent to the SocialCrawl API, which gathers public data from social networks, shops, maps, app stores, job boards, news and market quotes. Profiles, posts, comments and search results come back in one format whatever the platform. There are seven tools: `socialcrawl_find` picks an endpoint for a task described in plain words, `socialcrawl_endpoint` shows its contract, `socialcrawl_estimate` prices a call in advance, `socialcrawl_request` and `socialcrawl_collect` fetch data, and `socialcrawl_account` and `socialcrawl_manage` handle the balance, monitors and background jobs. Before a call the server validates parameters locally, so a malformed request spends no credits. Prices come in three kinds: per request tier (1, 5 or 10 credits), a flat rate for an endpoint, and metered, where a ceiling is held first and the difference is refunded. A preview with `dry_run=1` is free.
Who it is for. Marketers and analysts who want an agent to collect profiles, reviews and prices from dozens of platforms, and developers who build products on that data.
Good fit when
- You need to compare one creator's followers on TikTok, Instagram, YouTube and X, and the agent makes four calls from one request
- A team tracks brand reviews on Trustpilot and G2 and wants sentiment labels without a separate analysis step
- A buyer looks for the cheapest TV across Amazon, Walmart, Target and eBay
- Before a large export you need to know how many credits it will burn: `socialcrawl_estimate` gives the price and the worst case
- A researcher analyzes job listings and salary ranges from LinkedIn and Indeed
Not a fit when
- You need data from private accounts or direct messages: the service returns public data only
- You do not want to pay per request: after the 100 starter credits calls are paid
- You want your own scraper with full control over parsing, while here the data comes from a third-party API
- You need an export with a fixed final price: for metered endpoints it depends on the request
Example request
Compare @mkbhd's follower counts on TikTok, Instagram, YouTube and X and tell me first how many credits it costsMCP
- Transport
- stdio, http
- Authentication
- key is optional
Environment variables
SOCIALCRAWL_API_KEYsecret- API key with the sc_ prefix from the SocialCrawl dashboard. Without it only endpoint search, endpoint contracts and cost estimates work.
SOCIALCRAWL_LEGACY_TOOLS- The value 1 brings back the nine tool names from 1.x as thin wrappers.
Limitations
- The service is paid: a new account gets 100 credits without a card, then calls spend credits.
- The price of some endpoints is variable. News search, for example, costs 2 to 62 credits.
- The endpoint and price catalog inside the server is built at release time, so search and estimates can lag behind the live API. Data requests themselves go to the live API.
- The OAuth 2.1 mode for claude.ai connectors is still a preview and off by default. Connect with the key header.
- Version 2.0 has seven tools instead of eleven, and `SOCIALCRAWL_LEGACY_TOOLS=1` brings the old names back.
How to disable. Run `claude mcp remove socialcrawl` or delete the `socialcrawl` block from the client configuration file. Revoke the API key in the SocialCrawl dashboard.
FAQ
How do I connect SocialCrawl to Claude Code?
For a local run use `claude mcp add --scope user socialcrawl -e SOCIALCRAWL_API_KEY=sc_your_key -- npx -y socialcrawl-mcp`. The hosted server connects with `claude mcp add --scope user --transport http socialcrawl https://mcp.socialcrawl.dev/mcp --header "Authorization: Bearer sc_your_key"`.
Do I need a key to try it?
No. The `socialcrawl_find`, `socialcrawl_endpoint` and `socialcrawl_estimate` tools answer without a key and for free. The key is needed for the data requests. A new account holds 100 credits and needs no card.
How much does a call cost?
Base rates are 1 credit for a standard request, 5 for an advanced one and 10 for a premium one. A search across all sources costs 20 credits, and for metered endpoints `socialcrawl_estimate` gives the price with the worst case.
Which platforms are available?
Social networks (TikTok, Instagram, YouTube, X, LinkedIn, Reddit, Threads, Telegram and more), shops (Amazon, Walmart, eBay and more), Google Play and the App Store, Tripadvisor, Yelp, Trustpilot, G2, job boards, quotes, news and trends. `socialcrawl_find` without arguments returns the full list.
What changes when upgrading from 1.x?
Seven tools replace eleven. `SOCIALCRAWL_LEGACY_TOOLS=1` turns on the nine old names as wrappers during the move, and the mapping table is in the README.
Discussion
No questions or reviews yet.
Related
MCP serversOfficial Firecrawl MCP server: web search, page scraping to Markdown, site crawling and in-page actions
Official Apify MCP server: the agent runs ready-made Apify Store scrapers for social media, maps, search and shops
Metasearch for agents: one MCP server and HTTP API merge results from many search providers into unified JSON with deduplication.
An MCP server for the QuantumProxies proxy service that gives an agent scraping, search, site crawling and proxy endpoints
A research MCP server: searches the web, SEC, courts and patents, reads whole pages and checks that cited sources really exist.
A plugin that crawls sites and turns pages into Markdown with metadata: seven skills and an MCP server with nine tools
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| SocialCrawl MCP (this card) | yes | key is optional | 28 | |
| Apify MCP Server | yes | OAuth | 16,092 | runs ready-made scrapers from the Apify Store, while SocialCrawl returns one response format across platforms and prices each call in credits |
| Firecrawl MCP Server | yes | API key | 7,809 | turns web pages into Markdown, while SocialCrawl returns structured data from social networks, shops and job boards |
README in short
The page opens with connection options: the hosted HTTP server, the npm package and ready configuration blocks for Claude Code, Claude Desktop, Cursor, VS Code and Windsurf. Then come example requests, a table of the seven tools, the pricing rules and a description of the free labels the server adds to answers, such as comment sentiment.