grokipedia-py

Zero-dependency Python client for extracting structured content from Grokipedia pages.

Install

pip install grokipedia-py

uv pip install https://github.com/caentzminger/grokipedia-py.git
uv add "grokipedia-py @ git+https://github.com/caentzminger/grokipedia-py.git"

Quickstart

from grokipedia import from_url

page = from_url("https://grokipedia.com/page/13065923")

print(page.title)
print(page.slug)
print(page.intro_text)
print(page.infobox[:3])
print(page.lead_figure)
print([section.title for section in page.sections])
first_media = next(
    (
        subsection.media
        for section in page.sections
        for subsection in section.subsections
        if subsection.media
    ),
    [],
)
print(first_media[:1])
print(len(page.references))
print(page.links[:5])
print(page.metadata.keywords)
print(page.markdown[:500])
print(page.to_json(indent=2))

Parse raw HTML without network access:

from grokipedia import from_html

page = from_html(html, source_url="https://grokipedia.com/page/13065923")

Resolve a page from a title:

from grokipedia import page

page_obj = page('"Hello, World!" program')

Search for page URLs:

from grokipedia import search

results = search("hello world")
print(results[:5])

If this returns [], try:

results = search("hello world", respect_robots=False)

As of February 18, 2026, https://grokipedia.com/robots.txt disallows /api/, and /search is mostly client-rendered HTML.

Use class-based API with sitemap manifest caching:

from grokipedia import Grokipedia

wiki = Grokipedia(verbose=True)
result = wiki.page("The C Programming Language")
matches = wiki.search("programming language")

# Lazy sitemap lookup + cached child sitemap manifests.
url = wiki.find_page_url('"Hello, World!" program')
manifest = wiki.refresh_manifest()

Logging

The library uses Python's standard logging module (logger namespace: grokipedia).

import logging

logging.basicConfig(level=logging.INFO)
logging.getLogger("grokipedia").setLevel(logging.DEBUG)

Development & CI

This project stays runtime dependency-free (dependencies = []) and relies on the standard library for runtime behavior.

just setup
just fmt-py
just lint-py
just typecheck
just test
just ci

Robots behavior

from_url() enforces robots.txt by default.

• respect_robots=True (default): validate robots.txt before page fetch.
• search() first tries /api/full-text-search and falls back to /search HTML parsing.
• allow_robots_override=False (default): strict mode.
• if robots.txt is unavailable or malformed, the library fails closed with RobotsUnavailableError.
• if URL is disallowed, it raises RobotsDisallowedError.

You can bypass robots enforcement by setting either:

• respect_robots=False, or
• allow_robots_override=True

Data model

from_url() and from_html() return Page with:

• url
• slug
• title
• intro_text
• infobox (InfoboxField list for dt/dd fact rows)
• lead_figure (LeadFigure from the top figure image/caption when present)
• sections (Section tree with nested subsections; each section includes indexed media)
• references (Reference list)
• links (ordered unique links extracted from the main article)
• metadata (PageMetadata, including optional keywords)

Page also includes:

• lede_text (alias of intro_text)
• lead_media (alias of lead_figure)
• markdown
• to_dict() / to_json()
• from_dict() / from_json()

Exceptions

All library exceptions inherit from GrokipediaError.

• FetchError
• HttpStatusError
• PageNotFoundError
• RobotsUnavailableError
• RobotsDisallowedError
• ParseError