Download Yao Engine from the [download page](/download). Select your platform and architecture. ## Step 1: Launch Yao Engine Open Yao Engine. The main window shows Yao and Tai services.  > **Note:** On first launch, set App Root to an empty directory. Requires Yao Agents v1.6.1+. Click **Start**. Both services will show their PIDs and addresses once running.  ## Step 2: Open the welcome page Click **Visit** to open the welcome page in your browser, then click **Get Started**.  ## Step 3: Sign in **First login** — use the default root account: - **Email:** `root@yaoagents.com` - **Password:** `Yao123++`  > **Note:** Accounts are stored locally. To register additional users, configure SMTP first, otherwise they won't receive the verification code. ## Step 4: Configure models Open **Settings → Model Configuration** (avatar icon, bottom-left of sidebar).  | Slot | Used for | |------|----------| | Default | General tasks | | Heavy | Complex reasoning, long context | | Light | Fast, lightweight queries | | Vision | Image recognition | | Audio | Speech-to-text | | Embedding | Vector search, knowledge base | > **Note:** DeepSeek models do not support vision or audio. Assign a separate provider to those slots (e.g. Claude or GPT-4o for Vision, Whisper for Audio). Click **Add Provider** and enter your credentials.  > **Tip:** Add an image2 (OpenAI-compatible) provider for text-to-image and image editing. Agents will automatically dispatch image generation tasks when needed. Select your default model.  ## Step 5: Install the desktop client The desktop client is recommended for daily use. It works alongside the browser. Download from the [client download page](/download/yaoagents). Enter your server address (same as the browser URL above) and click **Connect**. Sign in the same way as you did in the browser.  > **Tip:** Right-click the tray icon to switch servers or access settings. ## What's next - [AI Experts](/docs/en-us/getting-started/ai-experts-intro) – Start working with built-in experts - [Connect a device](/docs/en-us/tai-link/connect-first-device) – Add machines via Tai Link