All the data you want. Many platforms. One API.

Every successful request costs exactly one API request, the same for every platform and endpoint, no matter how large the response. Failed requests (4xx or 5xx) are never charged. Each response includes an x-credits-remaining header so you can track usage live.

Never. API requests are a one time purchase, not a subscription, so nothing resets at the end of a month. Your balance carries forward indefinitely and is shared across every platform and endpoint.

A wide, growing set of endpoints across many platforms: Instagram, TikTok, YouTube, Facebook, LinkedIn, X, Spotify and Apple Music, plus the German sites Kleinanzeigen, 11880, Gelbe Seiten and ImmobilienScout24. You get profiles, posts, videos, comments, ads and listings, all as clean, consistently shaped JSON. The full current list with schemas lives in the docs.

None. Your only ceiling is your remaining balance, not a requests per second cap, so you can run large backfills or sudden spikes without asking for a quota. Just parallelise sensibly and retry the occasional transient error.

Every call fetches live data through rotating residential proxies and authenticates with a single API key you control. Create, rotate or revoke it anytime, and a revoked key stops working instantly. Keys are shown once and stored only as a hash, so even we cannot recover the raw value.

Yes. Besides the REST API there is a CLI, an MCP server and a Claude Skill, all hitting the same endpoints with the same key and returning identical JSON. The CLI and MCP server install from npm in one line, so you can prototype in your shell and ship the same call to production.

Never. Every endpoint runs login free from your side: send your API key, get JSON back. All the platform plumbing, tokens, private API calls, signatures and pagination, is handled for you, so you never create or babysit accounts on the target platform.

It is fetched live at request time, so you get the current state, not a cached snapshot. Two calls a minute apart can legitimately differ. If you need a time series, call on your own schedule and store each result yourself.

Failed requests are never charged, so you only pay for data you receive. Errors come back as a clean JSON envelope with a clear message and the right HTTP status: 400 malformed, 401 or 403 a key problem, 404 not found, 5xx a transient upstream issue worth a retry. The shape is predictable, so you can branch on the status code directly.

Yes, they are a first class use case. With no rate limits you can fan out hundreds of concurrent requests, bounded only by your balance. Cost is easy to estimate: one API request per successful call, so 100,000 records is 100,000 requests. For very large jobs, reach out and we will provision the account to run it smoothly.