> ## Documentation Index
> Fetch the complete documentation index at: https://docs.neuralcleave.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Deployment

> Run NeuralCleave as a systemd service, Docker container, or Tauri desktop app.

## Local (development)

```bash theme={null}
neuralcleave start
cd frontend && npm run dev
```

## systemd (Linux server)

Create `/etc/systemd/system/neuralcleave.service`:

```ini theme={null}
[Unit]
Description=NeuralCleave AI Gateway
After=network.target

[Service]
Type=simple
User=neuralcleave
ExecStart=/usr/local/bin/neuralcleave start
Restart=on-failure
RestartSec=5s
Environment=OPENAI_API_KEY=sk-...
Environment=TELEGRAM_BOT_TOKEN=123456:ABC...

[Install]
WantedBy=multi-user.target
```

```bash theme={null}
sudo systemctl daemon-reload
sudo systemctl enable neuralcleave
sudo systemctl start neuralcleave
```

## Docker

```dockerfile theme={null}
FROM python:3.12-slim

WORKDIR /app
COPY . .
RUN pip install -e ".[channels]"

EXPOSE 7432

CMD ["neuralcleave", "start"]
```

```bash theme={null}
docker build -t neuralcleave .
docker run -d \
  -p 7432:7432 \
  -v ~/.neuralcleave:/root/.neuralcleave \
  -e OPENAI_API_KEY=$OPENAI_API_KEY \
  -e TELEGRAM_BOT_TOKEN=$TELEGRAM_BOT_TOKEN \
  neuralcleave
```

### Docker Compose

```yaml theme={null}
version: "3.9"
services:
  gateway:
    build: .
    ports:
      - "7432:7432"
    volumes:
      - neuralcleave-data:/root/.neuralcleave
    environment:
      OPENAI_API_KEY: ${OPENAI_API_KEY}
      TELEGRAM_BOT_TOKEN: ${TELEGRAM_BOT_TOKEN}
    restart: unless-stopped

volumes:
  neuralcleave-data:
```

## Tauri desktop app

Build a self-contained desktop app (bundles both the gateway and the dashboard):

```bash theme={null}
cd frontend
npm run tauri build
```

The output is a platform-native installer (`.exe` on Windows, `.dmg` on macOS, `.AppImage` / `.deb` on Linux). The installer ships:

* The Next.js frontend (static export)
* The Python gateway as a Tauri sidecar binary
* A system tray icon with start/stop controls

## Reverse proxy (nginx)

If you expose the dashboard publicly:

```nginx theme={null}
server {
    listen 443 ssl;
    server_name ai.example.com;

    location /api/ {
        proxy_pass http://127.0.0.1:7432;
        proxy_http_version 1.1;
        proxy_set_header Upgrade $http_upgrade;
        proxy_set_header Connection "upgrade";
    }

    location / {
        proxy_pass http://127.0.0.1:3000;
    }
}
```

<Warning>
  Do not expose the gateway API publicly without adding authentication. The gateway has no built-in auth layer.
</Warning>

## Environment variables reference

| Variable             | Required                | Description        |
| -------------------- | ----------------------- | ------------------ |
| `OPENAI_API_KEY`     | If using OpenAI         | OpenAI API key     |
| `ANTHROPIC_API_KEY`  | If using Anthropic      | Anthropic API key  |
| `TELEGRAM_BOT_TOKEN` | If using Telegram       | Telegram bot token |
| `DISCORD_BOT_TOKEN`  | If using Discord        | Discord bot token  |
| `ELEVENLABS_API_KEY` | If using ElevenLabs TTS | ElevenLabs key     |