how to build workflow-use?

[sunch20]

I read the documentation and there are still some things I don't understand.
I want to use this workflow-use on my mac computer. Do I have to follow all the steps in the picture below?

[morugu]

@sunch20
You can use workflow-use just by running Setup workflow environment — no need to both steps.

If you want to record your own workflows, you have two options:

• Using the browser extension: Please build the extension locally.
• Using "Record your own workflow": This also allows you to record workflows without using the extension.

[balaboom123]

Qt. Do I have to follow all the steps in the picture below?
yes, you are correct

here is a more detailed version for you.

---

1. Install prerequisites (optional. depend on whether you have the tools)

1. Homebrew (if you don’t have it):

   /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"

2. Python 3.11+ and Node 16+:

   brew install python node

3. uv (for syncing your Python deps):

   brew install uv

---

2. Clone & build the browser-extension

# pick a folder, then:
git clone https://github.com/browser-use/workflow-use.git
cd workflow-use/extension

# install & build
npm install
npm run build

You should now have your built extension artifacts in extension/.output/chrome-mv3/

---

3. Set up the Python “workflows” environment

cd ../workflows

# 1) Generate & sync a venv from pyproject.toml + uv.lock
uv sync

# 2) Activate it
source .venv/bin/activate

# 3) Install Playwright browsers
playwright install chromium

# 4) Copy & edit your .env
cp .env.example .env
# → open .env in your editor and paste in your OPENAI_API_KEY

---

4. Verify everything

# List available commands
python cli.py --help

# Run the shipped example workflow to make sure it works:
python cli.py run-workflow examples/example.workflow.json 

[sunch20]

@balaboom123
Thank you for your reply, there is one more small question. Does the browser extension generated in the second step need to be installed on the browser?

[sunch20]

@morugu thank you for your reply.
there is one more question need your help. if I want use Deepseek, What content in the engineering code do I need to modify ？

[morugu]

@sunch20
You can use DeepSeek by initializing the model like this:

from langchain_deepseek import ChatDeepSeek

llm_instance = ChatDeepSeek(model='deepseek-chat')

Make sure to set your DEEPSEEK_API_KEY as an environment variable.

---

Alternatively, you can wait for this PR to be merged, which adds support for DeepSeek

[balaboom123]

@sunch20

@balaboom123 Thank you for your reply, there is one more small question. Does the browser extension generated in the second step need to be installed on the browser?

No, you don't need to. As the demo showed in readme.md or link here, you can see the user type the command in the terminal to run and it will pop up a browser (chromium) automatically. If you are interesting, you can check the technique of browser-use and playwright.