The Scrape API for
Clean Markdown & JSON
Turn any URL into clean markdown or schema-validated JSON in one request. DataBlue handles JavaScript rendering and proxy rotation for you, and the Firecrawl-compatible surface means migrating takes minutes, not a rewrite. Sign up and get 1,000 one-time credits to test it against real pages.
Any Webpage, Already Parsed.
A web scraping APItakes a URL and hands back its content as structured data, not raw HTML you have to fetch and parse yourself. That sounds simple until a site sits behind JavaScript rendering or a bot check. A scraper that worked Monday breaks silently by Friday, and you won't know until your data goes stale.
DataBlue removes that maintenance burden entirely. Send a URL, and it races a fast HTTP fetch against a full browser render, then returns whichever one comes back valid first. Ask for markdown, JSON, or any other supported format, and it arrives in the same request. There's no proxy pool to babysit, and no HTML-to-markdown converter to maintain on your side.
Because the surface is Firecrawl-compatible, teams evaluating a Firecrawl alternative usually just point an existing integration at DataBlue by swapping the base URL and the API key.
From URL to Structured Data.
One REST call to the Scrape endpoint returns the page, already parsed. SDK examples for crawl and search live on their own workflow pages.
# REST curl -X POST "https://api.datablue.dev/v1/scrape" \ -H "Authorization: Bearer wh_..." \ -H "Content-Type: application/json" \ -d '{ "url": "https://news.ycombinator.com", "formats": ["markdown", "links"] }'
{
"success": true,
"data": {
"markdown": "# Hacker News\n\n1. Show HN...",
"links": [...],
"metadata": {
"title": "Hacker News",
"status_code": 200,
"word_count": 1847
}
}
}Every Format You Need,
in One Call.
Request one format or several. DataBlue returns exactly what your pipeline needs, and you convert nothing yourself.
markdown
Clean, LLM-ready markdown with the boilerplate stripped out. This is the default format most pipelines reach for first.
html / rawHtml
Sanitized HTML for downstream parsing, or the untouched raw HTML exactly as the browser received it.
links
Every link found on the page, ready to seed a crawl or build a site map.
screenshot
A full page or viewport capture, handy for visual QA or for archiving what a page looked like at request time.
json (schema extraction)
Typed fields matching a JSON Schema you define. Ask for prices or release dates and get exactly that back.
metadata
Title and status code, bundled with every response so you never make a second call for basic context.
Built for Production, Priced for Scale.
Plenty of scrape APIs hand you HTML and call it a day. What sets DataBlue apart shows up over thousands of pages a day, not in the first request. Pricing is the clearest example. Every scrape carries a flat, transparent credit weight, with no hidden multiplier for JavaScript rendering or proxy use. Monthly plan credits reset each billing period, and top-up credits never expire.
On the technical side, DataBlue races an HTTP fetch against a full browser render and keeps whichever strategy wins per domain. JavaScript-heavy pages render correctly without you ever choosing a mode up front. Page actions and wait conditions give you control over what gets captured, and mobile device emulation is a single flag away. A request that doesn't return data doesn't cost you a credit either.
The API surface being Firecrawl-compatible means output is LLM-ready by default. Clean markdown drops straight into a RAG pipeline. A native MCP server lets agents like Claude and Cursor scrape pages directly, no glue code required. Wire it into your agent stack. Watch the first scrape come back structured.
What Teams Build With It.
A reliable scrape API quietly powers a surprising amount of software, including RAG pipelines and AI agents.
LLM & RAG ingestion
Turn any URL into clean markdown for a retrieval pipeline. No HTML parsing or boilerplate stripping code to maintain on your end.
AI agent web access
Give agents a single tool call to read a live webpage. The native MCP server lets Claude and Cursor scrape directly, no custom wiring needed.
Structured data extraction
Pull typed JSON, prices or specs, straight from a schema-validated scrape instead of writing a parser for every site.
Content & change monitoring
Re-scrape a page on a schedule and diff the markdown or JSON output. Catch price changes or content updates the moment they happen.
Firecrawl migration
Swap in DataBlue's Firecrawl-compatible scrape endpoint. Cut costs without rewriting your request or response handling.
Visual QA & archiving
Capture full-page screenshots alongside markdown or JSON. Keep a visual record of exactly what was scraped and when.
Your First Scrape in 60 Seconds.
No sales call and no contract to sign. Sign up, grab a key, and your first scrape is one request away.
Create a Free Account
Sign up and grab an API key instantly. You get 1,000 one-time signup credits to test real pages, no credit card required. Keys use the wh_ prefix and never expire.
Send Your First Scrape
Call /v1/scrape with a Bearer token and a URL. Choose your output with the formats field, markdown or schema JSON, whatever the job calls for.
Read the Output
Pull markdown or your schema fields straight off the response. No HTML to parse, no browser to manage. Ship it.
Scrape credit weights are visible before you run.
Start free with 1,000 one-time signup credits. Plans and endpoint weights come straight from the live catalog, so pricing stays straightforward even as that catalog changes. There are no hidden premium multipliers and no surprise fees. Monthly plans reset on schedule, and top-up credits stay available until you spend them.
Scrape API Questions.
What is a Scrape API?
A scrape API takes a URL and hands back the page's content as clean, structured data instead of raw HTML. You skip the browser automation and the proxy pool. There's no parser to patch when the layout changes. DataBlue's version renders JavaScript and returns markdown or JSON in a single request, with proxy rotation built in. Send a URL and see the structured response for yourself.
How do I convert a webpage to markdown with an API?
Send a POST request to /v1/scrape with formats: ["markdown"]. The response holds clean markdown in data.markdown. It's ready to drop straight into an LLM context window. Navigation bars and cookie banners get stripped automatically, so what's left is the actual page content. Point a RAG pipeline at it. Watch the manual cleanup step vanish.
Does the Scrape API handle JavaScript-rendered pages?
Yes. DataBlue runs a fast HTTP fetch and a full browser render at the same time, then keeps whichever one returns a valid result first. You never choose a rendering mode up front. Sites that need heavier handling get upgraded to a stealth rendering strategy automatically, and that choice is cached per domain. Try it against a JavaScript-heavy page. Watch the render race do its job.
Can I extract structured JSON instead of markdown?
Yes. Pass formats: ["json"] with a JSON Schema. The API returns typed fields matching that schema, not freeform text. Define whatever you need, like prices or release dates. The response comes back shaped exactly that way. The concept mirrors Firecrawl's schema extraction, so an existing schema usually works with a light edit. Drop in your schema. Check the output shape yourself.
Is this compatible with Firecrawl?
Yes. DataBlue's scrape and extract endpoints are shaped to match Firecrawl's, so migrating usually means swapping a base URL and an API key. You won't need to rewrite your integration. If you're evaluating a Firecrawl alternative, the request and response shapes will already look familiar. Point your existing client at DataBlue. Run one test call.
What output formats are supported?
Six formats in total, each returning a different shape of the same page. Request one or several in a single call through the formats array. Each format carries its own options too, like onlyMainContent for markdown and HTML, or fullPage for screenshots. Scroll to the breakdown below and pick what your pipeline actually needs.
Can the API click, scroll, or wait before scraping?
Yes, through page actions and options like wait_for and css_selector. That covers common cases like waiting out lazy-loaded content or dismissing a cookie banner before the page gets captured. For finer control, include_tags and exclude_tags narrow exactly what gets pulled into the response. Set the wait condition once and reuse it across similar pages.
How much does the Scrape API cost?
DataBlue charges a flat, transparent credit weight per scrape. There's no hidden multiplier for JavaScript rendering or extraction, so the price you see going in is the price you pay. Monthly plan credits reset every billing period, and top-up credits never expire. A failed request that returns no data doesn't cost a credit either. New accounts start with 1,000 one-time signup credits, enough to test real pages before spending a cent.
Can I scrape pages that need a mobile viewport?
Yes. Set mobile_device to true. The page then renders with a mobile viewport and user agent. That matters more than it sounds, since plenty of sites serve different markup to mobile visitors than to desktop ones. Flip the flag. Compare the two outputs side by side.
What happens if a page fails to load or blocks scraping?
DataBlue retries automatically through its managed proxy and rendering layer. If a request still fails to return valid data, it counts as a failure rather than a success. Failed requests don't consume a credit. You only pay for pages that actually come back.
Start Scraping in JSON.
Grab a free key and send your first request. Read back clean markdown or schema-based JSON from any URL, no card required, with 1,000 one-time signup credits to get you started.