SearchCans solves a two-step problem that every agentic or RAG application hits: find the right pages, then get their content into a format an LLM can actually consume. The SERP side returns structured JSON from Google and Bing, covering standard web search plus Shopping, Images, News, Videos and Short Videos, with country and language targeting on each request. The Reader side takes a URL — or a PDF or Office document — and returns clean Markdown or HTML, with a configurable JavaScript wait so heavy single-page applications have time to render before extraction.
The API is designed to be called in sequence with a single key. A typical quick-start does a search for a query, takes the first result URL and immediately posts it to the reader endpoint, printing Markdown that is ready to feed into a model. Because both halves draw on the same credit pool, there is no separate provisioning step when you move from prototyping to production.
The commercial model is deliberately different from most SERP APIs. There are no hourly rate limits; instead throughput is governed by parallel lanes, where one lane equals one concurrent in-flight request. At a three-second average latency, a single lane corresponds to roughly 1,200 requests per hour, so a five-lane plan is about 6,000. Eligible plans can be stacked so lanes accumulate in one account, and the Ultimate tier adds a dedicated cluster node with no queue. Credits are prepaid rather than subscription-based, run from $0.56 to $0.90 per thousand after discount depending on pack size, remain valid for six months, and 100 free credits are granted at signup without a card.
