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

# Quick Start

> Get CLaaS running with Local, Tinker, or Modal backends

# Quick Start

## Prerequisites

* **Python 3.11+**
* **[uv](https://docs.astral.sh/uv/)**, fast Python package manager
* **Docker** (for Local and Tinker Docker deployments)
* **GPU with >= 24 GB VRAM** (Local backend only)

```bash theme={null}
git clone https://github.com/kfallah/CLaaS.git
cd CLaaS
```

## Choose your backend

<Tabs>
  <Tab title="Local GPU">
    Runs SDPO training and vLLM inference on your own GPU. Requires >= 24 GB VRAM.

    ```bash theme={null}
    # Install dependencies
    uv sync --extra local

    # Docker setup (recommended)
    cd docker
    cp .env.local.example .env
    # Edit .env -- set TELEGRAM_BOT_TOKEN
    docker compose --profile local up --build
    ```

    The first run downloads Qwen3-8B (\~16 GB). Expect the vLLM health check to take 10–20 minutes on first start.

    <Card title="Full Local backend reference" icon="arrow-right" href="/training/local">
      Requirements, all config variables, services, and manual setup
    </Card>
  </Tab>

  <Tab title="Tinker SDK">
    Uses Tinker's hosted inference and training. **No GPU required.**

    ```bash theme={null}
    # Install dependencies
    uv sync --extra tinker

    # Docker setup
    cd docker
    cp .env.tinker.example .env.tinker
    # Edit .env.tinker -- set TELEGRAM_BOT_TOKEN + TINKER_API_KEY
    docker compose -f docker-compose.tinker.yml --env-file .env.tinker up --build
    ```

    <Card title="Full Tinker backend reference" icon="arrow-right" href="/training/tinker">
      Tinker config, model naming gotchas, and services
    </Card>
  </Tab>

  <Tab title="Modal Cloud (Coming Soon)">
    Runs distillation remotely on Modal (L40S GPUs). No local GPU needed.

    ```bash theme={null}
    # Install dependencies
    uv sync --extra local

    # Authenticate with Modal
    uv run modal token new

    # Deploy
    export HF_TOKEN=...
    export CLAAS_BASE_MODEL_ID=Qwen/Qwen3-8B
    uv run modal deploy -m claas.deploy
    ```

    The deployed app exposes the same API at `https://your-app--claas-distill-fastapi-app.modal.run`.

    <Card title="Full Modal backend reference" icon="arrow-right" href="/training/modal">
      Deployment, config, and LoRA storage on Modal Volumes
    </Card>
  </Tab>
</Tabs>

## Verify your setup

Once the stack is running, verify with:

```bash theme={null}
# Check inference endpoint
curl http://localhost:8000/v1/models -H "Authorization: Bearer sk-local"

# Check CLaaS API
curl http://localhost:8080/

# List LoRA adapters
curl http://localhost:8080/v1/lora
```