# Quickstart

> Create the starter configuration, point it at a model, run a first task, continue a session in the REPL, and open the dashboard.

This guide assumes Raw is installed. If not, follow [Install Raw](https://raw.tlelabs.com/docs/start/installation/) first.

1. **Create the starter configuration.**

   ```sh
   raw config init
   ```

   This writes `~/.config/raw/config.json` with a default agent named `raw`. It does not overwrite an existing file.

2. **Point the model at a real model.**

   Open the file and find the `models` block. The starter entry looks like this:

   ```json
   {
     "default_agent": "raw",
     "models": {
       "local": {
         "provider": "ollama",
         "method": "openai-chat-completions",
         "model_id": "YOUR_INSTALLED_MODEL",
         "base_url": "http://127.0.0.1:11434/v1"
       }
     }
   }
   ```

   Replace `YOUR_INSTALLED_MODEL` with the exact model ID your server reports. For a hosted model, add another entry to `models` instead. Keep the API key in an environment variable and reference it with `api_key_env`:

   ```json
   {
     "models": {
       "flash": {
         "provider": "deepseek",
         "method": "openai-chat-completions",
         "model_id": "deepseek-flash",
         "base_url": "https://api.deepseek.com",
         "api_key_env": "DEEPSEEK_API_KEY",
         "context_window_tokens": 1048576
       }
     }
   }
   ```

   Then point an agent at the new alias by setting its `model` field to `flash`. See [Models and agents](https://raw.tlelabs.com/docs/guides/models-and-agents/) for every field.

3. **Run a first task.**

   ```sh
   raw "Explain the tests in this repository"
   ```

   Raw runs the task with the default agent, prints the answer to stdout, and writes status lines to stderr. When the run ends, it prints a copyable resume command. To choose an agent explicitly, use `raw --agent raw "task"`.

4. **Open the interactive REPL.**

   ```sh
   raw
   ```

   The REPL keeps one saved session. Type a task at the `❯` prompt. Use `/compact` to summarize older context, `/clear` to start a new saved session, `/stats` to show usage, and `/exit` to leave. Press Ctrl-C to cancel active work.

5. **Continue the last session.**

   ```sh
   raw --continue "Follow up on that change"
   ```

   `--continue` resumes the most recent unexpired session for the current workspace. To resume a specific session, use `raw --resume ID "task"`. To list sessions, run `raw sessions`.

6. **Open the dashboard.**

   ```sh
   raw dashboard
   ```

   Raw starts a local server at `http://127.0.0.1:8787`, opens your browser, and prints an authenticated link. Keep the process running while you use the dashboard. Setup and saved history work even before the model connection is configured.

## Next steps

- [Core concepts](https://raw.tlelabs.com/docs/start/concepts/) explains how models, agents, tools and skills fit together.
- [Permissions](https://raw.tlelabs.com/docs/guides/permissions/) explains what a tool call can do on your machine.
- [Dashboard](https://raw.tlelabs.com/docs/guides/dashboard/) describes the browser app.