# English Style Converter

A SvelteKit web app that converts English text into various styles and tones using an LLM.

## Quick Start (Docker)

```bash
# Option 1: Use default model (llama3)
docker compose up

# Option 2: Choose a different model
OLLAMA_MODEL=gemma2 docker compose up
```

- **App:** http://localhost:3000
- **Ollama API:** http://localhost:11434

First startup pulls the model from Ollama, which may take a few minutes depending on model size and your connection. Subsequent starts are instant (model is cached in a Docker volume).

To change the model later, edit `.env.docker` and run:

```bash
docker compose down
docker compose up --build
```

## Local Development (without Docker)

Prerequisites: [Ollama](https://ollama.ai) running locally with a model pulled.

```bash
# Install dependencies
npm install

# Copy env config (defaults to Ollama at localhost:11434)
cp .env.example .env

# Start dev server
npm run dev
```

## Configuration

| Variable | Default | Description |
|----------|---------|-------------|
| `OPENAI_BASE_URL` | `http://localhost:11434/v1` | OpenAI-compatible API endpoint |
| `OPENAI_API_KEY` | `ollama` | API key (use `ollama` for local Ollama) |
| `OPENAI_MODEL` | `llama3` | Model to use |
| `PORT` | `3000` | App port (Docker/adapter-node only) |

For Docker, set `OLLAMA_MODEL` in `.env.docker` — it controls both the model Ollama pulls and the model the app requests.

## Styles

6 categories, 25 styles: Sarcastic, Formal, British (Polite, Formal, Witty, Gentlemanly, Upper Class, Royal, Victorian, Downton Abbey), American (New Yorker, AAVE, Southern, Redneck), Pirate, Shakespearean, Gen Z, Game of Thrones (King's Landing, Wildlings, Winterfell), and Newspeak (Orwellian).

## Testing

```bash
npm test
```