brave-search-python-client 0.2.10

🦁 Brave Search Python Client supporting Web, Image, News and Video search.

🦁 Brave Search Python Client

Brave Search Python Client supports Web, Image, News and Video search.

Use Cases:

1. Add it to your Python code so your users find what they are looking for.
2. Add it to the AI application you are building so your LLM has access to up-to-date information in the web.
3. Use the integrated CLI in your shell scripts - it provides search results as JSON.

Setup

Installation is as simple as:

uv add brave-search-python-client               # add dependency to your project

If you don't have uv installed follow these instructions.

In case you don't have an API key for Brave Search yet, sign-up here - it's free up to 2.000 requests per month.

See the code sample below and explore the reference documentation on how to use the Brave Search Python Client as part of your own code.

If you just want to try out the client without having to write code you can use the integrated CLI:

export BRAVE_SEARCH_API_KEY=YOUR_API_KEY         # replace YOUR_API_KEY
uvx brave-search-python-client --help            # show available commands
uvx brave-search-python-client web "hello world" # search for hello world

If you still prefer pip over the modern and fast package manager uv, you can install like this:

pip install brave-search-python-client          # add dependency to your project

Highlights

• Modern async Python client for Web, Image, Video and News search powered by Brave Search
• Various Examples:
  • Streamlit web application deployed on Streamlit Community Cloud
  • Jupyter notebook
  • Simple Python script
• Thorough validation of both requests and responses (powered by Pydantic)
• Complete reference documenation on Read the Docs
• 100% test coverage including unit and E2E tests (reported on Codecov)
• Matrix tested with Python 3.11, 3.12 and 3.13 to ensure compatibility (powered by Nox)
• 100% compliant with modern linting and formatting standards (powered by Ruff)
• 100% up-to-date dependencies (monitored by Renovate)
• A-grade code quality in security, maintainability, and reliability with 0 technical debt and 0 codesmell (verified by SonarQube)
• 1-liner for installation and execution of command line interface (CLI) via uv(x) or Docker
• Setup for developing inside a devcontainer included (supports VSCode and GitHub Codespaces)
• Later: MCP server to connect Brave Search with Claude Desktop and other MCP clients

Usage Examples

Streamlit App

Try it out! - Show the code

Minimal Python Script:

import asyncio
import json
import os

from dotenv import load_dotenv

from brave_search_python_client import (
    BraveSearch,
    CountryCode,
    ImagesSearchRequest,
    LanguageCode,
    NewsSearchRequest,
    VideosSearchRequest,
    WebSearchRequest,
)

# Load .env file and get Brave Search API key from environment
load_dotenv()
api_key = os.getenv("BRAVE_SEARCH_API_KEY")
if not api_key:
    raise Exception("BRAVE_SEARCH_API_KEY not found in environment")


async def search():
    """Run various searches using the Brave Search Python Client (see https://brave-search-python-client.readthedocs.io/en/latest/reference_index.html)"""

    # Initialize the Brave Search Python client, using the API key from the environment
    bs = BraveSearch()

    # Perform a web search
    response = await bs.web(WebSearchRequest(q="jupyter"))

    # Print results as JSON
    print("# Web search")
    print("## JSON response")
    print(json.dumps(response.model_dump(), indent=2))

    # Iterate over web hits and render links in markdown
    print("## Iterate and render")
    for result in response.web.results if response.web else []:
        print(f"[{result.title}]({result.url})")

    # Advanced search with parameters
    response = await bs.web(
        WebSearchRequest(
            q="python programming",
            country=CountryCode.DE,
            search_lang=LanguageCode.DE,
        )
    )
    print("# Advanced search results")
    for result in response.web.results if response.web else []:
        print(f"[{result.title}]({result.url})")

    # Search and render images
    print("# Images")
    response = await bs.images(ImagesSearchRequest(q="cute cats"))
    for image in response.results if response.results else []:
        print(f"![{image.source}]({image.url})")

    # Search and render videos
    print("# Videos")
    response = await bs.videos(VideosSearchRequest(q="singularity is close"))
    for video in response.results if response.results else []:
        print(f"![{video.title}]({video.url})")

    # Search and render news
    print("# News")
    response = await bs.news(NewsSearchRequest(q="AI"))
    for item in response.results if response.results else []:
        print(f"![{item.title}]({item.url})")


# Run the async search function
# Alternatively use await search() from an async function
asyncio.run(search())

Show script code - Read the reference documentation

Jupyter Notebook

Show notebook code

Command Line Interface (CLI)

Run with uvx

Add Brave Search API key to the environment

export BRAVE_SEARCH_API_KEY=YOUR_API_KEY

Show available commands:

uvx brave-search-python-client --help

Search the web for "hello world":

uvx brave-search-python-client web "hello world"

Show options for web search

uvx brave-search-python-client web --help

Search images:

uvx brave-search-python-client images "hello world"

Show options for image search

uvx brave-search-python-client images --help

Search videos:

uvx brave-search-python-client videos "hello world"

Show options for videos search

uvx brave-search-python-client videos --help

Search news:

uvx brave-search-python-client news "hello world"

Show options for news search

uvx brave-search-python-client news --help

Run with Docker

Note: Replace YOUR_BRAVE_SEARCH_API_KEY with your API key in the following examples.

Show available commands:

docker run helmuthva/brave-search-python-client --help

Search the web:

docker run --env BRAVE_SEARCH_API_KEY=YOUR_BRAVE_SEARCH_API_KEY helmuthva/brave-search-python-client web "hello world"

Show options for web search

docker run helmuthva/brave-search-python-client web --help

Search images:

docker run --env BRAVE_SEARCH_API_KEY=YOUR_BRAVE_SEARCH_API_KEY helmuthva/brave-search-python-client images "hello world"

Show options for image search

docker run helmuthva/brave-search-python-client images --help

Search videos:

docker run --env BRAVE_SEARCH_API_KEY=YOUR_BRAVE_SEARCH_API_KEY helmuthva/brave-search-python-client videos "hello world"

Show options for video search

docker run helmuthva/brave-search-python-client videos --help

Search news:

docker run --env BRAVE_SEARCH_API_KEY=YOUR_BRAVE_SEARCH_API_KEY helmuthva/brave-search-python-client news "hello world"

Show options for news search

docker run helmuthva/brave-search-python-client news --help

Extra: MCP Server

TK

Contributing

Please read our Contributing Guidelines for how to setup your development environment, and guidance for making pull requests.

Resources

• API
• MCP Specification and SDKs

Star History