Core APIs

Extract

Fetch known URLs and get back clean, readable content without navigation, ads, or boilerplate.

Overview

Extract reads one or more URLs you already have and returns cleaned content. Pass a schema to also get structured JSON per page (see Structured extraction).

POST/extract

Request

curl https://api.ohsee.tech/extract \
  -H "Authorization: Bearer ohsee-<your-key>" \
  -H "Content-Type: application/json" \
  -d '{
    "urls": ["https://modelcontextprotocol.io/docs"],
    "format": "markdown"
  }'

Parameters

Field	Type	Default	Description
`urls`required	string[]	-	1 to 20 URLs to extract.
`format`	string	markdown	Output format: `markdown` or `text`.
`include_favicon`	bool	false	Include each source's favicon URL.
`schema`	object	null	If set, runs structured extraction on each page. Charged at structure rates.
`cache`	string	auto	Cache behavior (`auto`, `force`, `bypass`); ignored when a schema is provided.
`cache_ttl`	int	3600	Cache lifetime in seconds (60 to 86400).

Response

Field	Type	Description
`results`	object[]	Successful pages, each with `url`, `raw_content`, optional `favicon`, and `structured` when a schema is used.
`failed_results`	object[]	Pages that could not be fetched, each with `url` and `error`.
`response_time`	float	Server processing time in seconds.
`request_id`	string	Unique ID for this request.
`credits_charged`	int	Credits deducted for this request.
`cache_status`	string \| null	Cache result for this request.

200 OK

{
  "results": [
    {
      "url": "https://modelcontextprotocol.io/docs",
      "raw_content": "# Model Context Protocol ...",
      "favicon": null
    }
  ],
  "failed_results": [],
  "response_time": 0.88,
  "request_id": "a91c33d0",
  "credits_charged": 1,
  "cache_status": "miss"
}

Credits

1 credit per 5 successfully extracted URLs (rounded up). Failed URLs are not charged. With a schema, pages are charged at structured-extraction rates.

Structured extraction