AI Agents in Container

コンテナ内の AI エージェント

kasm デスクトップ

docker run \
  --name browser-use \
  --shm-size=4096m \
  -w /mnt \
  -p 8080:6901 \
  -e VNC_PW=password \
  -id kasmweb/ubuntu-noble-desktop:1.16.1
  
user: kasm_user
password: password

browser-use

• https://github.com/browser-use/browser-use.git

Python の最新バージョンをインストール

docker exec -u root -it browser-use bash
visudo
kasm-user ALL=(ALL) NOPASSWD: ALL
passwd kasm-user
passwd root
exit

docker exec -it browser-use bash
sudo apt update -y
sudo apt install -y python3-venv git xvfb python3-full

web-ui

• https://github.com/browser-use/web-ui.git

# Start in Desktop directory
cd /home/kasm-user/Desktop

# Remove any existing web-ui directory if it exists
sudo rm -rf web-ui

# Clone the repository
git clone https://github.com/browser-use/web-ui.git
cd web-ui

# Create and set up virtual environment in your home directory instead
python3 -m venv ~/web-ui-venv
source ~/web-ui-venv/bin/activate

# Install UV
curl -LsSf https://github.com/astral-sh/uv/releases/latest/download/uv-installer.sh | sh
source $HOME/.local/bin/env

# Verify environment
whoami
uv --version
python3 --version

# Install requirements using the virtual environment in home directory
uv pip install -r requirements.txt

# Install Playwright dependencies
playwright install-deps 
playwright install

# Setup environment file
cp .env.example .env

# Run the application with Xvfb
xvfb-run -a --server-args="-screen 0 1280x1024x24" python webui.py --ip 0.0.0.0 --port 7788

# Access the application at:
# http://localhost:7788


# Remove existing venv
deactivate
cd /home/kasm-user/Desktop
sudo rm -rf web-ui

OLLAMA をインストール

curl -fsSL https://ollama.com/install.sh | sh

ollama serve &

http://localhost:11434

ollama pull deepseek-r1:8b

ollama pull deepseek-r1:8b  # I used this for docker

ollama run deepseek-r1:8b

ollama list

ollama stop deepseek-r1:8b

ollama rm deepseek-r1:8b

MAC OS

• Python の最新バージョンをインストール

brew install python3

git clone https://github.com/browser-use/web-ui.git

# Install UV Package https://docs.astral.sh/uv/getting-started/installation/

cd web-ui

python3 -m venv .venv

source .venv/bin/activate

curl -LsSf https://github.com/astral-sh/uv/releases/latest/download/uv-installer.sh | sh

source $HOME/.local/bin/env

uv --version

python3 --version

uv pip install -r requirements.txt

playwright install-deps 

playwright install

cp .env.example .env

python webui.py --ip 127.0.0.1 --port 7788

http://localhost:7788

MAC OS に OLLAMA をインストール

https://ollama.com/download/mac

ollama pull deepseek-r1:8b  # I used this for Mac OS

ollama run deepseek-r1:8b

Vscode Insiders

• https://code.visualstudio.com/insiders/

Google AI Studio（API Key）

• https://aistudio.google.com/apikey

MCP サーバー

• https://github.com/modelcontextprotocol/servers

• https://github.com/punkpeye/awesome-mcp-servers?tab=readme-ov-file#file-systems

• https://smithery.ai/

• これによりインターネットにアクセスできるため、最新のドキュメントを取得し、それに応じてコードを更新できます

git clone https://github.com/modelcontextprotocol/servers.git
cd servers
docker build -t mcp/puppeteer -f src/puppeteer/Dockerfile .

• vscode Insider の設定

{
    "mcp": {

      "inputs": [],
      "servers": {
        "puppeteer": {
          "command": "docker",
          "args": ["run", "-i", "--rm", "--init", "-e", "DOCKER_CONTAINER=true", "mcp/puppeteer"]
        }
      }
    }
    
  }

• Claude Desktop の設定

{
  "mcpServers": {
    "puppeteer": {
      "command": "docker",
      "args": ["run", "-i", "--rm", "--init", "-e", "DOCKER_CONTAINER=true", "mcp/puppeteer"]
    }
  }
}

• Cursor

{
  "mcpServers": {
    "puppeteer": {
      "command": "docker",
      "args": ["run", "-i", "--rm", "--init", "-e", "DOCKER_CONTAINER=true", "mcp/puppeteer"]
    }
  }
}

• Codium Windsurf

{
  "mcpServers": {
    "puppeteer": {
      "command": "docker",
      "args": ["run", "-i", "--rm", "--init", "-e", "DOCKER_CONTAINER=true", "mcp/puppeteer"]
    }
  }
}