Social Fetch
Get API key
All use cases

Creator intelligence API

Seed creator lists, resolve the same person across networks, compare engagement on recent posts, and pull text for topic tagging — without maintaining a scraper per platform or normalizing five different JSON shapes in your warehouse.

Try in playgroundView pricing

What usually breaks first

Most creator intelligence pipelines fail on schema drift and billing surprises long before they fail on "missing a platform." The notes below are the failure modes we hear about after a DIY scraper or marketplace API is already in production.

Creator intelligence products need four capabilities that home-grown scrapers usually ship one at a time: finding new creators, linking handles across networks, comparing engagement with the same math, and reading enough recent content to know what someone actually posts about.

TikTok reports views and likes on each video. YouTube puts subscriber counts under different keys. Instagram paginates differently and hides half the fields in GraphQL fragments. If you write adapters for each network, every new feature waits on whichever platform broke this week. Your users still want a creator card that reflects follower counts from yesterday, not from the last manual CSV export.

Agency teams keep client rosters as rows of handles per platform. SaaS products store creator graphs with one node per network and manual edges for "same person." Both need public profile data on demand: follower totals for shortlists, recent uploads for cadence checks, captions or transcripts when you classify topics. Social Fetch returns the same envelope — data, meta, lookupStatus, metrics, profile — on every network so your database schema stays flat and your scoring code does not fork.

8 routes for this job

These are the routes we see in live creator intelligence integrations — TikTok, YouTube, Instagram, and X in this list. Method, path, and query params match the public OpenAPI spec; when we add a field to docs, it ships in responses.

  • GET /v1/tiktok/users/search

    Keyword search to seed discovery pipelines when you expand a vertical or backfill a new niche.

  • GET /v1/tiktok/profiles/{handle}

    TikTok identity card: followers, bio, avatar, verification. Anchor node when you merge cross-platform records.

  • GET /v1/youtube/channel

    YouTube channel by handle, channel ID, or URL. Subscriber count lands in the same metrics object as TikTok and Instagram.

  • GET /v1/instagram/profiles/{handle}

    Instagram profile with bio links — often the fastest signal for which YouTube or TikTok handle belongs to the same person.

  • GET /v1/twitter/profiles/{handle}

    X/Twitter profile using the same path-parameter pattern as Instagram. Fan out in parallel with other networks.

  • GET /v1/tiktok/profiles/{handle}/videos

    Recent TikTok uploads with per-item engagement. Compute median views or posting frequency in your own aggregation layer.

  • GET /v1/instagram/profiles/{handle}/posts

    Paginated Instagram posts for benchmarking likes and comment counts against TikTok from the same creator row.

  • GET /v1/tiktok/videos/transcript

    Spoken text from a video URL when captions are thin and you need keywords beyond hashtags for theme scoring.

OpenAPI reference →

How teams wire this

A common path starts with “discover and seed candidates.” Your cron cadence, warehouse schema, and alert thresholds will differ — treat the steps as ordering hints, not a checklist you must copy verbatim.

  1. 1

    Discover and seed candidates

    Start with TikTok user search (or platform search endpoints for your vertical). Persist each handle with the query that found it so you can re-run the same list after a campaign or model refresh. Filter client-side on follower floors before you spend credits on full enrichment.

  2. 2

    Resolve cross-platform identities

    For each candidate, call profile endpoints in parallel — TikTok, YouTube channel, Instagram, X. Match declared links in bios first, then your own fuzzy rules on display names. Each response carries its own data.lookupStatus: found, not_found, and private are completed lookups you can store; retry only on lookup_failed.

  3. 3

    Merge into a creator record

    Write one warehouse row per person with child rows per platform. Pull metrics from the shared envelope instead of mapping subscriberCount on YouTube to followerCount on TikTok by hand. Log meta.creditsCharged per call so agency ops can split cost per client import.

  4. 4

    Benchmark engagement

    Fetch recent videos or posts for creators above your cutoff. Calculate engagement rate, views per post, or days between uploads in SQL or pandas — the API returns paginated items with per-post numbers, not a pre-baked score you cannot audit.

  5. 5

    Extract content themes

    Concatenate captions from post endpoints, or call transcript routes when speech carries the signal. Run your own classifier or keyword extractor on the raw text; we return strings, not opaque topic labels tied to a black-box model.

  6. 6

    Refresh before deliverables

    Re-pull profiles and content on a cron before QBRs, campaign reviews, or ranking model retrains. Credits are prepaid and do not expire between quarterly bursts, so a quiet creator graph between agency reporting cycles does not run up a subscription.

Example: TikTok profile for a creator graph node

Swap YOUR_API_KEY for a key from the dashboard. The playground pre-fills auth if you open the same path there — useful before you paste this into a worker or CI job.

Request
Example: TikTok profile for a creator graph node

Why intelligence products build on Social Fetch

  • One JSON envelope across 13+ platforms — your ETL maps metrics and profile once, not once per network adapter.
  • lookupStatus on every completed call so ranking pipelines store not_found and private accounts explicitly instead of treating HTTP 200 as proof of data.
  • Discovery search, profile resolution, post history, and transcripts bill from the same credit balance and API key.
  • No published rate cap on metered routes: parallelize thousands of profile lookups during a backfill within your balance.
  • meta.creditsCharged on each response — allocate spend per handle, per batch job, or per agency client in your own ledger.
  • 100 free credits to run real handles through your scoring code before you buy credits for a full import.
  • When a platform changes its HTML upstream, your client keeps parsing the same response shape.
  • Operated by Social Freak Ltd — a registered UK company when procurement asks who runs the API.

Run a creator intelligence lookup on free credits

You get 100 credits on signup — enough to walk through the curl above against live data and inspect lookupStatus, pagination, and meta.creditsCharged. If the JSON lands cleanly in your pipeline, buy a credit pack once; balances do not expire on a subscription clock.

Open playgroundSee pricing