Skip to content

Latest commit

 

History

History
88 lines (45 loc) · 2.84 KB

File metadata and controls

88 lines (45 loc) · 2.84 KB

🚀 How to Boot Up & Shut Down Your Sovereign AI Agent

Navigate to your main project folder in your Ubuntu terminal before starting: cd ~/Desktop/Project-G/linuxech-sovereign-agent (adjust path if needed)

🟢 STARTUP SEQUENCE

Terminal 1: Wake up the AI Brain (Ollama)

In your first terminal window, start the local LLM server:

ollama serve

(Note: If it says "bind: address already in use", that just means Ollama is already running in the background as a system service. You can close this terminal if that happens!)

Terminal 2: Start the Python Backend (The Hands)

Open a second terminal window, navigate to your project folder, and start the FastAPI server:

cd linuxech-sovereign-agent/backend source venv/bin/activate python3 agent_backend.py

Wait until you see the message: Application startup complete. It is now listening on port 8000.

Terminal 3: Start the React Dashboard (The Face)

Open a third terminal window, navigate to your project folder, and start the Vite frontend:

cd linuxech-sovereign-agent/frontend npm run dev

You will see a green Vite logo and a local URL (usually http://localhost:5173).

🌐 ACCESSING THE APP

Hold the Ctrl key on your keyboard and click the http://localhost:5173 link in Terminal 3 (or copy-paste it into your web browser).

The dark-mode Linuxech HITL Pipeline dashboard will appear!

Upload your resumes, lock in your job roles, and click "HUNT JOBS" to watch the AI go to work.

⚠️ A quick reminder about your i5 CPU: When the Python backend triggers browser-use to auto-apply, a Chromium browser will physically pop open. Because your CPU is doing heavy AI math to read the screen, your laptop fans will spin up, and the browser might sit still for 15-30 seconds between clicks. Do not close it! Just let the CPU do its work.

🛑 GRACEFUL SHUTDOWN SEQUENCE

When you are done applying for jobs, it is important to close the servers properly to free up your 16GB of RAM and stop your CPU from running background AI tasks.

Step 1: Stop the React Frontend

Go to Terminal 3.

Press Ctrl + C on your keyboard.

This stops the Vite development server.

Step 2: Stop the Python Backend

Go to Terminal 2.

Press Ctrl + C on your keyboard. You should see a message saying Shutting down.

To exit the Python virtual environment safely, type:

deactivate

Step 3: Put the AI to Sleep (Ollama)

Go to Terminal 1.

Press Ctrl + C to stop the Ollama server and unload the Llama/Qwen models from your system memory.

Step 4: Clean Up Zombie Browsers (Optional but Recommended)

Sometimes, if you interrupt the AI while it is applying for a job, a "headless" (invisible) Chromium browser might get stuck running in the background, secretly eating your RAM.

To ensure all AI-controlled browsers are forcefully closed, run this command in any terminal:

pkill -f chromium pkill -f playwright

(If it outputs nothing, you are perfectly clean!)