Turn any URL into clean JSON for your LLM pipeline.
Scrapiq strips the boilerplate from any web page and returns structured data — perfect for RAG ingestion, fine-tuning dataset generation, and any LLM workflow that needs clean source input instead of HTML soup.
# pip install httpx
import httpx
resp = httpx.post("https://scrapiq.io/v1/extract", json={
"url": "https://news.ycombinator.com",
"format": "markdown",
})
print(resp.json()["content"][:200])Three formats. One endpoint.
Pick what your pipeline needs. Get it back in < 200ms for typical pages.
Clean text
Boilerplate stripped: nav, ads, footer, scripts, cookies banners. Just the article.
{ "format": "text" }
Markdown with links
Same as text, but with hyperlinks preserved. Drop into any RAG splitter directly.
{ "format": "markdown" }
Structured JSON
Pass a JSON schema. Scrapiq extracts the matching fields. No LLM in the loop.
{ "format": "json", "schema": {...} }
Live playground
Real call to scrapiq.io/v1/extract. Try a URL you control.
// Click "Run extraction" to see the response here.Why Scrapiq?
No signup
v0.1 is public. Send a request, get a response. No key, no account, no quota tracking.
Self-hostable
git clone + pip install -e . and you have your own instance on any VPS. No vendor lock-in.
Built for LLM pipelines
FastAPI · httpx · trafilatura · MIT. ~200ms p50 latency for typical pages. Sensible caching built in.
Honest
Pre-revenue, open-source. No fake testimonials, no inflated numbers. What you see is what runs.
Ship your RAG pipeline today.
One endpoint. Three formats. MIT license.