# Running the App on Windows

## Requirements

- Windows 10 or 11
- Python **3.10, 3.11, or 3.12**
  - Download from [python.org/downloads](https://www.python.org/downloads/)
  - During installation **check "Add Python to PATH"**
- An internet connection (for API calls and initial package download)

---

## Step 1 — Download the project

Copy the project folder to your Windows machine (USB, zip, or any file transfer).
Place it somewhere easy to find, e.g. `C:\Users\YourName\narjis_project`.

---

## Step 2 — Run the setup script

Open **File Explorer**, navigate to the project folder, then double-click:

```
setup_windows.bat
```

What it does:
- Checks that Python is installed
- Creates an isolated virtual environment (`.venv` folder)
- Installs all required packages:
  - `streamlit` — the web UI framework
  - `openai` — ChatGPT / OpenAI API
  - `anthropic` — Claude API
  - `google-genai` — Gemini API
  - `requests` — NVIDIA streaming API
- Optionally installs `transformers` + `torch` if you want to run local models (needs a GPU and ~10 GB disk space)

When it finishes you will see **"Setup complete!"**

---

## Step 3 — Start the app

Double-click:

```
run_app.bat
```

A terminal window will open and you will see:

```
You can now view your Streamlit app in your browser.
Local URL: http://localhost:8501
```

Open your browser and go to **http://localhost:8501**

To stop the app, click the terminal window and press **Ctrl + C**.

---

## Using the app

### 1. Choose a provider (sidebar)

| Provider | What you need |
|---|---|
| ChatGPT (OpenAI) | OpenAI API key from platform.openai.com |
| Claude (Anthropic) | Anthropic API key from console.anthropic.com |
| Gemini (Google) | Google AI API key from aistudio.google.com |
| NVIDIA | NVIDIA API key from build.nvidia.com |
| Local (GPU) | A GPU with enough VRAM (no API key needed) |

Paste your API key in the password field — it is **never saved to disk**.

### 2. Pick a model

Each provider shows its available models in a dropdown.

### 3. Answer research questions

- The main area shows three **Move** panels (Move 1, 2, 3).
- Select a Move, then choose a specific question from the dropdown.
- Type your answer and click **➕ Add answer**.
- Repeat for as many questions as you like across all three Moves.
- Use the **✕** button next to any answer to remove it.

### 4. Generate

Click **✨ Generate introduction**. The app combines your instruction prompt and all answers into one structured prompt and sends it to the selected model. The result appears below.

---

## Troubleshooting

**"Python was not found"**
Re-install Python and make sure "Add Python to PATH" is checked.

**"pip install failed"**
Run the setup script again. If your network requires a proxy, configure it in Windows network settings first.

**App opens but API call fails**
- Double-check your API key (no extra spaces).
- Make sure the key has access to the selected model.
- Some models (e.g. GPT-4.1, Claude Opus 4.7) require a paid tier.

**Blank page or port conflict**
Edit `run_app.bat` and change `8501` to another port, e.g.:
```
.venv\Scripts\streamlit.exe run app.py --server.port 8502
```
Then open http://localhost:8502.