What is a social media API?
A social media API returns structured JSON for profiles, posts, comments, search, or media metadata over HTTP instead of scraped HTML. The phrase covers three different things — official partner kits, public-data REST APIs, and scraper marketplaces — with different access models and integration shapes.
Three shapes people mean
Official platform APIs (TikTok Kit, Instagram Graph, YouTube Data API) require app review and OAuth for owner-scoped data — publishing, your channel's analytics, your page's posts. Wrong tool for looking up someone else's public profile.
Public-data REST APIs take an API key and return public profiles, posts, and search across networks with a consistent envelope. Social Fetch: one key, one shape, no OAuth — every lookup targets public data.
Scraper runtimes (Apify-style) accept a job, crawl asynchronously, and return a dataset later — different from a synchronous GET. Useful for custom crawl logic; awkward when you need inline answers.
What you get back from Social Fetch
Public-data REST: one API key across 21 platforms and 165 operations — TikTok through web extraction. Every response uses { data, meta } with requestId, creditsCharged, and version.
Profile calls return data.lookupStatus (found, not_found, private) on HTTP 200. Route via /platforms, /docs, /tools, or POST /v1/ask for plain-English routing.
How to pick one
Publishing, DMs, or owner analytics → official kit. Public enrichment across networks with one schema → public-data REST API. One-off custom crawl no endpoint covers → scraper runtime, accepting job latency and schema drift.
Compare: sync vs async, schema stability, metering model, and whether you get a correlation ID on failures.
How this compares across the industry
Official kits optimize for trust and scoped tokens — unusable for data you don't own. Scraper marketplaces optimize coverage — job-based, schemas drift with markup. Public-data REST sits between: narrower than general scraping, but synchronous and typed.
Common mistakes
Treating 200 with lookupStatus not_found/private as a transport error — it bills normally; retry loops waste credits.
Assuming OAuth is required everywhere — only for owner-scoped official kits; public-data APIs use API keys.
Ignoring metering until production — Social Fetch has no hard RPS cap on paid routes (~500 concurrent guidance); other providers cap differently.
What the term doesn't mean
Not a single spec or standard — due diligence on any product matters. Not automatically real-time or access to private content (DMs, deleted posts) in any category.
FAQ
What is a social media API?
HTTP JSON for profiles, posts, comments, or media — covering official kits, public-data REST APIs, or scraper marketplaces with different shapes.
Is a social media API the same as a scraper?
Not always. Some are typed REST APIs over maintained infrastructure; others expose raw async crawl jobs. Compare latency, schema stability, and requestId support.
Do I need OAuth for a social media API?
For owner-scoped official kit access, yes. Public-data APIs typically need only an API key.
How much does a social media API cost?
Varies. Social Fetch: PAYG packs from $14/1k credits, 100 free on signup, no required subscription.
Which platforms does Social Fetch cover?
21 platforms under one key: TikTok, Instagram, YouTube, X/Twitter, LinkedIn, Facebook, Reddit, Threads, Telegram, GitHub, Spotify, Rumble, Bluesky, Hacker News, Truth Social, Twitch, Ad Libraries, and web extraction.
What does a typical response look like?
{ data, meta }. data holds payload and lookupStatus on lookups; meta always has requestId, creditsCharged, version.
Can I query with plain English instead of picking an endpoint?
Yes — POST /v1/ask or nl_ask_post. Routing is 0 credits; the nested lookup bills normally.
Is there a rate limit I need to plan around?
No hard RPS cap on paid routes; credit balance is the limit. ~500 concurrent requests recommended for reliability.