ScrapeCreators API

Use ScrapeCreators when a workflow needs structured public social or ad-library data at a scale that ordinary web search cannot provide.

Runtime selection

Task skills describe each request as an environment-neutral operation containing provider, method, path, and optional query or body. Execute that same operation through the first available runtime:

1. GooseWorks MCP available: inspect and call the live call_data_provider tool. Pass the operation fields directly. This is the preferred path in ChatGPT, Cowork, and any client without a terminal. It uses the first-party GooseWorks ScrapeCreators proxy and managed key; it does not route through Orthogonal.
2. Local GooseWorks CLI available: translate the operation into gooseworks call <provider> <path>, adding the method, query, or body options represented in the operation. The CLI uses the same first-party proxy.
3. User-owned ScrapeCreators key available: call https://api.scrapecreators.com<path> with SCRAPECREATORS_API_KEY in the x-api-key header, using the operation's method, query, and JSON body.
4. No runtime available: ask the user to connect GooseWorks MCP, sign in through the local GooseWorks CLI, or provide their own ScrapeCreators key. Do not pretend the operation ran.

Prefer MCP whenever call_data_provider is registered, even if a shell might also exist. Never ask for a separate provider key when the managed MCP or CLI path is available. Never print, store, or return either credential.

Example operation: