What if your own language model could search the web and answer real-time questions? Just like ChatGPT, but running locally on your system, inside your app, where you have full control?
In this project, we’ll build exactly that using:
- 🧠 Ollama – to run powerful language models like
gemma3on your local machine. - 🔗 LangChain – to interface with LLMs from Python.
- 🌐 Bright Data MCP – to fetch live web content using real-world scraping tools. Bright Data MCP GitHub Repo
- 🎨 Streamlit – to wrap everything in a clean, interactive user interface.
Before running the app, make sure you have the following:
- ✅ WSL 2 + Miniforge, here's how to install it: https://youtu.be/luM5kwH6tjQ OR Linux + any other Conda type of software.
- ✅ Ollama installed and a model (e.g.,
gemma3) pulled and serving. (check the video tutorial below for insturctions) - ✅ An API key from Bright Data, make sure that at least one Web Unlocker zone is configured.
- ✅ Add the following line of code to the end of your .bashrc file in WSL. Replace with the Brigth Data API Key you copied.
export BRD_API_KEY=<your-key>
If you're not sure how to install the requirements, or what .bashrc is, or if you'd like to build this entire application step by step alongside me, please checkout my video tutorial on YouTube:
- Connect a local LLM to real-time web data
- Use MCP tools for scraping websites, X posts, and LinkedIn profiles
- Build a Streamlit GUI with optional context fields and memory
- Understand how MCP gives models access to live external resources
Get 10$ of Bright Data credits using the promo link (goes a long way!):
https://brdta.com/pythonsimplified_mcp
Just for perspective. I've featured Bright Data tools in a few tutorials already, and so far it cost me $6 bucks.
# Clone the repo
git clone https://github.com/MariyaSha/simple_mcp_app.git
cd simple_mcp_app
# Create working environment
conda create -n mcp_env python=3.12
conda activate mcp_env
conda install nodejs
# Make sure you've updated your .bashrc file (check Prerequisites)
npx @brightdata/mcp API_TOKEN=$BRD_API_KEY
# Install Python dependencies
pip install -r requirements.txt
cd simple_mcp_app
# OR cd advanced_mcp_app
# Run application
streamlit run mcp_app.py