Social Fetch?

NOT REALLYA prototype of the interface is possible, but the durable product depends on difficult systems beyond the UI. Build a safer adjacent workflow instead.

BUILD FIRST

A non-sensitive or read-only version of the core experience
A manual review step wherever correctness or trust matters
A narrow companion tool that removes a real piece of work

LEAVE FOR LATER

Regulated decisions, payments, or irreversible actions
Claims of correctness, compliance, or uptime that the MVP cannot support
Sensitive integrations until security, support, and auditability are ready

RESEARCH NOTES

What this tool actually does.

Checked 2026-09-06

Social Fetch (socialfetch.dev) is a hosted REST API and MCP server that scrapes and normalizes social media data. It exposes 184 endpoints covering 23 platforms including Instagram, TikTok, X/Twitter, and YouTube. Data includes profiles, posts, search results, video transcripts, and engagement metrics, returned in one consistent JSON envelope. Authentication uses an x-api-key header (sfk_...); base URL is https://api.socialfetch.dev. It offers a TypeScript SDK with typed methods and Result-based error handling, plus integrations for Cursor, Claude Code, n8n, Apify, and x402. The hosted MCP server (https://api.socialfetch.dev/mcp) is a Remote Streamable HTTP server over OAuth or x-api-key, exposing each operation as a typed tool plus docs_search, docs_read, and nl_ask_post helpers. Agent-facing docs include /llms.txt, /agents.txt, /llms.json, and per-platform /llms-{platform}.txt files, w

SOURCES

RESEARCHED BUILD PROMPT

Give this to your coding agent.

Build a focused replacement for Social Fetch, a social media data normalization API. Core workflow: a developer signs up, gets an API key, and calls a REST endpoint like GET /v1/tiktok/profiles/{handle} to receive normalized profile, post, transcript, or engagement data in a consistent JSON envelope across multiple social platforms. Every response includes a meta field with lookupStatus and creditsCharged. Implement: (1) a REST API with OpenAPI schema at /openapi.json, key auth via x-api-key header, and per-platform routes; (2) a TypeScript SDK with typed client methods and Result-based error handling; (3) a hosted MCP server exposing the same operations as typed tools and supporting OAuth; (4) agent-friendly documentation with /llms.txt, /agents.txt, and /llms.json. Use a scraping layer (headless browser pool or third-party scrapers), a normalization layer mapping platform-specific fields to one schema, a credit-billing meter, and a documentation site. For the MCP server use the Remote Streamable HTTP spec with FastMCP or similar. Do not copy the original brand name, logo, color scheme, or site layout. Out of scope: real multi-platform anti-bot evasion, production-scale proxy rotation, payment processing UI, and live platform access for all 23 platforms — stub at least 3 platforms realistically and design the registry to add more. Remove the nl_ask_post plain-English helper from scope. Deliver a monorepo with the API server, SDK, MCP server, docs generator, and a README covering setup, environment variables, how to add a new platform adapter, credit accounting behavior, and known limitations (rate limits, platform ToS risk, stale-cache strategy). Mention Social Fetch's workflow explicitly in the README as the reference product being rebuilt, but note it is a clean-room implementation with no copied code, branding, or UI.

Visit the official product page