Search the web with advanced filters and LLM-friendly output formats!

from search_ai import search

results = search('What is the best LLM in 2025?')

for result in results:
    print(result.title)

Output:

LLM Leaderboard 2025 - Verified AI Rankings
Best LLM Models 2025: Top 10 AI Models Ranked & Compared
Top 7 LLMs Ranked in 2025: GPT-4o, Gemini, Claude & More
10 Best Large Language Models (LLMs) in 2025 - Beebom
Best Large Language Models (LLMs) of 2025 - TechRadar
Top 9 Large Language Models as of September 2025 | Shakudo
Top 15 LLMs in 2025: Best Large Language Models
Top LLMs in 2025: Comparing Claude, Gemini, and GPT-4 LLaMA
Best LLM Models 2025 - Complete Guide to Top 7 AI Language Models
15 Best LLM Models in 2025 | Top AI Language Models Ranked

Install

$ pip install search-ai-core

Using filters

from search_ai import search, Filters

search_filters = Filters(
    in_title="python",              # Only include results with "python" in the title
    tlds=[".edu", ".org"],          # Restrict results to .edu and .org domains
    https_only=True,                # Only include websites that support HTTPS
    exclude_sites='quora.com',      # Exclude results from quora.com
    exclude_filetypes='pdf'         # Exclude PDF documents from results
)

results = search(filters=search_filters)
for result in results:
    print(result.title)

Output:

Welcome to Python.org
Python Tutorial - W3Schools
Python (programming language) - Wikipedia
Learn Python - Free Interactive Python Tutorial
CS50's Introduction to Programming with Python | Harvard University
Real Python: Python Tutorials
Python for Everybody Specialization - Coursera
scikit-learn: machine learning in Python — scikit-learn 1.6.1 ...
Table Of Contents - Learn Python the Hard Way
Python Institute - PROGRAM YOUR FUTURE

Regional targeting

from search_ai import search, Filters, Regions

search_filters = Filters(region=Regions.JAPAN)

results = search('Python', filters=search_filters)

for result in results:
    print(result.title)

Output:

Welcome to Python.org
python.jp: プログラミング言語 Python 総合情報サイト
【入門】Pythonとは｜活用事例やメリット、できること、学習方法 ...
ゼロからのPython入門講座 - python.jp
Pythonの開発環境を用意しよう！（Windows） - Progate
Python - Wikipedia
プログラミング言語のPythonとは？その特徴と活用方法 - 発注ナビ
Python試験・資格、データ分析試験・資格を運営する一般社団法人 ...
Pythonの導入方法｜ソフトの利用方法 - 東京経済大学
Pythonとは？開発に役立つ使い方、トレンド記事やtips - Qiita

Markdown & JSON formats

Once extracted, you can retrieve the results in either Markdown or JSON format for further processing.

If the extend argument is set to True, the content of the result's websites will also be included in the output. To achieve this functionality, SearchAI uses Playwright to load and extract content from websites. In addition to extracting the main content of a page, SearchAI also tries to find metadata on pages, such as an author name and twitter handle.

Getting results in markdown (example):

SearchResults.markdown(
    extend=False,           # Set to True to fetch and include page content
    content_length=1000,    # Limit the length of extracted content
    ignore_links=False,     # Exclude hyperlinks in the content
    ignore_images=True,     # Exclude images from the content
    only_page_content=False # If True, omits metadata from the output
)

Getting results in json (example):

SearchResults.json(
    extend=False,           # Set to True to fetch and include page content
    content_length=1000,    # Limit the length of extracted content
    ignore_links=False,     # Exclude hyperlinks in the content
    ignore_images=True,     # Exclude images from the content
)

Using proxies

If you'd like to use proxies, you can create a proxy object using Proxy and pass it into either search or async_search.

from search_ai import Proxy, search

proxy = Proxy(
    protocol="[protocol]",
    host="[host]",
    port=9999,
    username="optional username",
    password="optional password"
)

search('query', proxy=proxy)

Async support

SearchAI also supports Asyncio! Instead of using search, use async_search. The async version will return an AsyncSearchResults which will contain multiple instances of AsyncSearchResult.

from search_ai import async_search

results = await async_search(...)
await results.json(extend=True)

All filters

You can narrow down searches by including filters like so:

Filters(
    sites="example.com",
    tlds=[".edu", ".gov"],
    filetype="pdf",
    exclude_sites=["facebook.com", "twitter.com"],
    in_title="python",
    not_in_url=["login", "signup"]
)

Here is a complete list of all the filters in SearchAI:

Filter	Description	Example (one)	Example (many)
region	Only show results from specific regions	Regions.US_ENGLISH
time_span	Timespan for the search	Timespans.PAST_WEEK
sites	Only show results from specific domains	"example.com"	["example.com", "another.com"]
tlds	Only show results from specific top-level domains (e.g. .gov, .edu)	".edu"	[".edu", ".gov"]
filetype	Only show documents of a specific file type (only one allowed)	"pdf"
https_only	Only show websites that support HTTPS	True
exclude_sites	Exclude results from specific domains	"facebook.com"	["facebook.com", "twitter.com"]
exclude_tlds	Exclude results from specific top-level domains	".xyz"	[".xyz", ".ru"]
exclude_filetypes	Exclude documents with specific file types	"doc"	["doc", "xls"]
exclude_https	Exclude HTTPS pages	True
any_keywords	Require at least one word anywhere in the page	"python"	["python", "django"]
all_keywords	Require all of these words somewhere in the page	"ai"	["ai", "ml", "nlp"]
exact_phrases	Include results with exact phrases	"machine learning"	["deep learning", "language model"]
exclude_all_keywords	Exclude pages containing certain words	"ads"	["ads", "tracking"]
exclude_exact_phrases	Exclude results with exact phrases	"click here"	["click here", "buy now"]
in_title	Require specific words in the title	"resume"	["resume", "portfolio"]
in_url	Require specific words in the URL	"blog"	["blog", "tutorial"]
in_text	Require specific words in the page text	"case study"	["case study", "example"]
not_in_title	Exclude pages with specific words in the title	"login"	["login", "signup"]
not_in_url	Exclude pages with specific words in the URL	"register"	["register", "checkout"]
not_in_text	Exclude pages with specific words in the page text	"error"	["error", "404"]

Search Configuration Options

The search and async_search functions have the following parameters that you can use to optimize your searches with:

Parameter	Type	Description	Default
query	str	The search query string.	""
filters	Filters | None	Optional Filters object to narrow search results.	None
count	int	Number of results to return.	10
offset	int	Number of results to skip at the beginning.	0
proxy	Proxy | None	Optional Proxy object to route requests through a proxy.	None