🦁 Brave Search Python Client

Tip

📚 Online documentation - 📖 PDF Manual

---

The Brave Search Python Client provides Web, Image, News, and Video search capabilities.

Use Cases:

1. Integrate into your Python code to help users find what they're looking for.
2. Add to your AI applications to give LLMs access to current web information.
3. Use the built-in CLI in shell scripts to get search results in JSON format.

Overview

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

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

Obtain your Brave Search API key by signing up here - the free tier includes 2,000 requests per month. For guidance on how to integrate the Brave Search Python client into your code base check out the examples below and explore the reference documentation. 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 web "hello world" # search for hello world

All advanced search options of Brave Search are supported by the client and in the CLI:

# Find all German content about AI added in the last 24 hours
uvx brave-search-python-client web --country=DE --search-lang=de --units=metric --freshness=pd ai

The CLI provides extensive help:

uvx brave-search-python-client --help            # all CLI commands
uvx brave-search-python-client web --help        # all options for web search
uvx brave-search-python-client images --help     # all options image search
uvx brave-search-python-client videos --help     # all options video search
uvx brave-search-python-client news --help       # all options news search

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:

"""
Example script demonstrating the usage of the Brave Search Python Client.

For web, image, video and news search.
"""

import asyncio
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:
    msg = "BRAVE_SEARCH_API_KEY not found in environment"
    raise ValueError(msg)


async def search() -> None:
    """Run various searches using the Brave Search Python Client (see https://brave-search-python-client.readthedocs.io/en/latest/reference.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

    # Iterate over web hits and render links in markdown
    for _result in response.web.results if response.web else []:
        pass

    # Advanced search with parameters
    response = await bs.web(
        WebSearchRequest(
            q="python programming",
            country=CountryCode.DE,
            search_lang=LanguageCode.DE,
        ),
    )
    for _result in response.web.results if response.web else []:
        pass

    # Search and render images
    response = await bs.images(ImagesSearchRequest(q="cute cats"))
    for _image in response.results or []:
        pass

    # Search and render videos
    response = await bs.videos(VideosSearchRequest(q="singularity is close"))
    for _video in response.results or []:
        pass

    # Search and render news
    response = await bs.news(NewsSearchRequest(q="AI"))
    for _item in response.results or []:
        pass


# 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 (to come)

Further Reading

• Check out the reference with detailed documentation of public classes and functions.
• Our release notes provide a complete log of recent improvements and changes.
• In case you want to help us improve the Brave Search Python Client: The contribution guidelines explain how to setup your development environment and create pull requests.

Star History