scrapix 0.1.4

Scrapix API

Scrapix — Python SDK

Official Python client for Scrapix — an AI-powered web scraping and data extraction API. Fetch pages, collect URLs, crawl sites, and extract structured data from a single function call, with JS rendering, proxy rotation, captcha solving, and LLM-backed extraction handled server-side.

---

Table of contents

• Installation
• Authentication
• Quick start
• Endpoints
  • scrape — single page
  • collect — URLs from a page
  • crawl — collect + scrape
  • extract — structured data & summaries
• Common input options
• Output formats
• Error handling
• Custom host / advanced configuration
• Links
• License

---

Installation

pip install scrapix

Requires Python 3.9 or newer.

Authentication

All requests require an API key. Get one at scrapix.promptcloud.com and either pass it explicitly or load it from an environment variable:

import os
from scrapix import create_client

api = create_client(os.environ["SCRAPIX_API_KEY"])

Quick start

from scrapix import create_client, ScrapeInput

api = create_client("YOUR_API_KEY")

result = api.scrape(ScrapeInput(url="https://example.com"))
print(result.data.data)        # page content (HTML by default)
print(result.status_code)      # upstream HTTP status
print(result.credits_used)     # credits consumed by this call

create_client(api_key) targets https://api-scrapix.promptcloud.com by default. Override with create_client(api_key, host="https://your-host").

---

Endpoints

scrape — single page

Fetch one URL and return its content in the requested format.

from scrapix import create_client, ScrapeInput

api = create_client("YOUR_API_KEY")

result = api.scrape(ScrapeInput(
    url="https://example.com",
    output_format="markdown",   # html | text | markdown | docx | pdf | base64
    render=True,                # execute JS before returning
))

print(result.data.format)       # "markdown"
print(result.data.data)         # rendered markdown

collect — URLs from a page

Extract outgoing links from a page. Useful as a first step before crawling.

from scrapix import create_client, CollectInput

result = api.collect(CollectInput(
    url="https://example.com",
    urls_limit=100,
    include_paths=r"^/blog/",       # regex filter — keep only matching paths
    exclude_paths=r"/tag/|/author/",
    include_sitemap_urls=True,      # also harvest URLs from sitemap.xml
))

# result.links.data is a serialized string in result.links.format ("json" by default)
print(result.links.data)

crawl — collect + scrape

Collect URLs from a starting page and scrape each one in a single call.

from scrapix import create_client, CrawlInput

result = api.crawl(CrawlInput(
    url="https://books.toscrape.com/",
    urls_limit=10,                  # 1–25
    output_format="markdown",
    include_sitemap_urls=False,
))

for page in result.responses:
    print(page.url, "->", len(page.data.data), "chars")

extract — structured data & summaries

Run an LLM over a page to answer a natural-language query or extract structured data.

from scrapix import create_client, ExtractInput

# Free-form answer
result = api.extract(ExtractInput(
    url="https://example.com",
    query="What is this website about? Answer in one sentence.",
))
print(result.result)

# Structured extraction — provide a JSON Schema
from scrapix import StructuredOutputSchema

result = api.extract(ExtractInput(
    url="https://books.toscrape.com/",
    query="Extract the first 5 books on this page.",
    structured_schema=StructuredOutputSchema(
        query="Extract the first 5 books.",
        schema={
            "type": "object",
            "properties": {
                "books": {
                    "type": "array",
                    "items": {
                        "type": "object",
                        "properties": {
                            "title": {"type": "string"},
                            "price": {"type": "string"},
                        },
                    },
                }
            },
        },
    ),
))
print(result.structured_output.data)   # serialized in result.structured_output.format

---

Common input options

Every input model accepts the following fields in addition to its endpoint-specific ones:

Field	Type	Default	Description
url	str	—	Target URL (required).
timeout	int	40	Request timeout in seconds (10–1000).
max_retries	int	5	Max retries before the request fails.
render	bool	False	Execute JavaScript before capturing the page.
premium_proxies	bool	False	Route through premium residential/mobile proxies.
auto_proxy	bool	False	Escalate automatically: normal proxies → premium → render fallback.
use_captcha_solver	bool	False	Solve captchas when encountered.
use_cache	bool	True	Serve from the fresh-crawl cache when available.

Output formats

Format	Where it applies	Notes
html	scrape, crawl	Default.
text	scrape, crawl	Plain-text extraction.
markdown	scrape, crawl	HTML → Markdown conversion.
docx / pdf	scrape, crawl	Binary content, base64-encoded.
base64	scrape, crawl	Raw HTML as base64.
json (default)	collect, extract.structured_output	Other structured formats also allowed.
xml/yaml/toml	collect, extract.structured_output

Error handling

API errors raise scrapix.ApiException (and its subclasses BadRequestException, UnauthorizedException, NotFoundException, ServiceException, …):

from scrapix import create_client, ScrapeInput
from scrapix.exceptions import ApiException

try:
    result = api.scrape(ScrapeInput(url="https://example.com"))
except ApiException as e:
    print(f"HTTP {e.status}: {e.reason}")
    print(e.body)          # raw error body
    print(e.headers)       # response headers

Non-HTTP failures (network, timeout) raise the underlying urllib3 exception.

Custom host / advanced configuration

create_client wraps the common "Configuration + ApiClient + APIServicesApi" setup. When you need interceptors, per-request timeouts, or connection pooling, drop down to the raw client:

import scrapix

config = scrapix.Configuration(
    host="https://api-scrapix.promptcloud.com",
    api_key={"APIKeyHeader": "YOUR_API_KEY"},
)
config.retries = 3

with scrapix.ApiClient(config) as client:
    api = scrapix.APIServicesApi(client)
    result = api.scrape(
        scrapix.ScrapeInput(url="https://example.com"),
        _request_timeout=60,
    )

Every endpoint method accepts per-call overrides via _request_timeout, _headers, and _host_index. See the API reference for the full list.

Links

• Home: https://scrapix.promptcloud.com
• API reference: https://docs-scrapix.promptcloud.com/api-reference/scrapix-api
• Docs: https://docs-scrapix.promptcloud.com
• Source: https://github.com/promptcloud/scrapix-sdk/tree/main/python
• Issues: https://github.com/promptcloud/scrapix-sdk/issues
• Changelog: https://github.com/promptcloud/scrapix-sdk/releases

License

Released under the terms of the LICENSE file shipped with this package.