{ "openapi": "3.0.3", "info": { "title": "HTML Parser \u2014 Ksty.ch", "version": "1.0", "description": "Parse an HTML document into structured content \u2014 metadata, Open Graph, headings, links, images and readable text.\n\n# HTML Parser API\n\nTurns a raw HTML document into clean, structured JSON. Give it the HTML directly with `html`, or hand it a `url` and it fetches the page for you, and it returns the pieces you normally have to write a scraper for \u2014 the title and meta description, Open Graph and Twitter card tags, the canonical URL and favicon, every heading, link and image, and the readable plain-text body with a word count and reading-time estimate.\n\nParsing runs locally against a real HTML parser, so malformed markup, unclosed tags and messy real-world pages are handled the same way a browser would tolerate them \u2014 you get structure back, not an error.\n\n## Two ways to call it\n\nSupply the markup directly:\n\n```\nPOST /api/v1/html-parser\nX-Api-Key: ksty_test_...\nContent-Type: application/json\n\n{\"html\": \"Hello

Hi

World

\"}\n```\n\n\u2026or let the API fetch the page (relative links are resolved against the fetched URL automatically):\n\n```\n{\"url\": \"https://example.com\"}\n```\n\n```json\n{\n \"title\": \"Hello\",\n \"lang\": \"\",\n \"description\": \"\",\n \"canonical\": \"\",\n \"favicon\": \"\",\n \"meta\": {},\n \"opengraph\": {},\n \"twitter\": {},\n \"headings\": { \"h1\": [\"Hi\"] },\n \"links\": [],\n \"images\": [],\n \"text\": \"Hi World\",\n \"word_count\": 2,\n \"reading_time_min\": 1\n}\n```\n\n## Endpoints\n\nThe root endpoint returns everything. Three focused endpoints return only one section each, so you pay for and transfer just the part you need:\n\n- `html-parser` \u2014 the full parse (all sections above).\n- `html-parser.metadata` \u2014 title, language, description, canonical, favicon, and the raw `meta` / `opengraph` / `twitter` maps. Ideal for building link previews.\n- `html-parser.links` \u2014 every `` as `{url, text}`, plus every `` as `{src, alt}`, with relative URLs resolved to absolute.\n- `html-parser.text` \u2014 the readable plain-text body only, with `word_count` and `reading_time_min`.\n\n## Parameters\n\n- `html` \u2014 the HTML document to parse. Either this or `url` is required; if both are sent, `html` wins.\n- `url` \u2014 an `http(s)` URL to fetch and parse. The fetch is a single GET with an 8-second timeout and a 3 MB size cap; URLs that resolve to private or reserved IP addresses are refused (SSRF protection).\n- `base_url` \u2014 a base used to resolve relative links and images to absolute URLs. Defaults to `url` when you fetch, or the document's own `` when present.\n\n## Link resolution\n\nRelative links (`/about`, `../pricing`, `logo.png`) are turned into absolute URLs whenever a base is available \u2014 from `base_url`, the fetched `url`, or a `` in the document, in that order. When no base can be determined the original relative value is returned unchanged.\n\n## Readable text\n\n`text` is the visible body text with `

Hi World

" }, "url": { "type": "string", "description": "An http(s) URL to fetch and parse instead of passing html", "example": null } } } } } } } } }, "components": { "securitySchemes": { "ApiKeyAuth": { "type": "apiKey", "in": "header", "name": "X-Api-Key" } }, "schemas": { "Error": { "type": "object", "properties": { "error": { "type": "object", "properties": { "code": { "type": "string", "description": "Stable machine-readable code \u2014 branch on this, never on the message." }, "message": { "type": "string" }, "request_id": { "type": "string" } } } } } } } }