문서

Sign up, copy your API key, and make your first search call in about five minutes.

How to send your API key, manage keys, and read rate limit headers.

POST /v1/search — any service in the catalog (search engines, ecommerce & social, agent search) with one request and response format.

POST /v1/ai-search — cited results and, where the source writes one, an answer: from agent search APIs (DeepSeek, Yuanbao, Tavily, Exa) or ChatGPT's own answers. Automatic failover; ChatGPT runs as an async job.

POST /v1/fetch — get the cleaned main content of any public web page as Markdown or plain text.

POST /v1/batch — up to 5 searches in one request. Each one is billed like a single search, and one failure does not block the others.

GET /v1/services and GET /v1/engines — the search catalog with per-service prices, supported parameters and live status.

GET /v1/searches/{id} — retrieve the stored result of a past search for 30 days.

GET /v1/account — check your credit balance and this month's usage programmatically.

The error envelope and every error code, with its HTTP status and what to do about it.

Give Claude Code, Cursor, Claude Desktop or any MCP client web search and page fetching through SerpHubAPI.

Get search results and page content as Markdown in one call for RAG pipelines and agent tools.

How credits work, what each call costs, top-up tiers, and the four billing rules.

Getting receipts and invoices for your top-ups, and setting your billing details.

What we store about your API calls, for how long, and how to request deletion.

Where the data comes from and the rules for using SerpHubAPI.