andrewjamesford/browser-use

A browser-use demonstration - In this agent example using multiple LLMs we can see how a LLM can navigate a browser and retrieve information for users.

Agent Ollama R1 & Claude 3.5 sonnet & gpt-4o browser-use demo

This demo app shows how you can use 3 different LLM models to perform browser navigation and return the search results of comparing API prices.

Getting started

Clone the repo, then install uv the python dependency and everything else tool (if you haven't already).

Create a .env file with the following variables and appropriate keys.

OPENAI_API_KEY=
ANTHROPIC_API_KEY=

Installation

Install via uv venv

uv venv --python 3.11

Activate it

# For Mac/Linux:
source .venv/bin/activate

# For Windows:
.venv\Scripts\activate

Install dependencies

uv pip install browser-use
playwright install

Running agents

To run the agent make sure you have the appropriate keys in an .env file. If using Ollama install the ollama application and pull the appropriate model you wish to use.

For example:

python ./agent-claude35-sonnet.py

The call will return a JSON response and generate a gif file. To do this it will use playwright to open a chromium browser and perform navigation of websites to get the appropriate information.