First-Time Setup
After registering and signing in, spend 2 minutes on the configuration below to get the full AI coding experience. Every setting can be changed later on the "Settings" page.
Recommended Setup Flow
Verify your sign-in state: open "Settings → Profile" and confirm your username and email are displayed correctly. The login token is managed by the local kernel and renewed automatically — no repeated sign-ins needed in normal use.
Choose a default model: open "Settings → AI Models" and pick a model in the "Default model" dropdown. LingXizhi natively supports 15+ mainstream models including Kimi K3, GLM-5.2, and DeepSeek-V4, covering both official direct connections and channels such as Volcano Engine. Your choice applies to chat, completion, and RUSH tasks at once.
Check your subscription and usage: open "Settings → Subscription & Usage" to see your current plan and today's Agent run usage (Free: 10/day, Pro: 1000/day, Ultimate: unlimited). Usage resets automatically every day — see Plans & Quota.
(Optional) Bind your own API Key: if you already have a key from a model provider, bind it in settings (BYOK, Bring Your Own Key). You are billed by the provider with zero platform markup, and the key is stored with AES-256 encryption. See BYOK.
(Optional) Adjust appearance and rules: switch between dark/light themes and adjust font size in "Settings → Appearance"; toggle built-in rules as needed in "Settings → AI Rules".
The model list comes from the cloud: the IDE does not hardcode a model list locally — available models are fetched by the kernel from the server in real time. If the "AI Models" dropdown is empty, the network or the kernel is usually not ready yet; wait a moment or restart the IDE.
The Kernel and Port 3721
LingXizhi IDE's AI capabilities are provided by a built-in Rust kernel. The kernel runs in the background and listens only on the loopback address 127.0.0.1:
- Default port 3721. If that port is occupied, the kernel automatically tries the next ports in sequence (up to 10) and writes the actual port back to the config — no manual handling required.
- It binds only to
127.0.0.1and is not exposed externally, so there is generally no need to allow it through the firewall. - All local API calls between the IDE and the kernel (sign-in, model list, subscription queries, RUSH tasks, etc.) go through this port.
To check whether the kernel is running, run in a terminal:
curl http://127.0.0.1:3721/health
A normal response means the kernel is ready. If the connection fails, restart the IDE and the kernel will be brought up automatically.
Do not kill the kernel process manually: AI chat, completion, quota queries, and more all depend on the local kernel. Force-terminating the kernel process makes these features unavailable; restarting the IDE restores them.
Next Steps
- Migrating from VS Code or Cursor? See Migrate from VS Code to import your keybindings and settings in one click.
- Learn about daily Agent runs and reset rules for each plan: Plans & Quota.
- Start your first task: press Ctrl + Shift + R to launch a RUSH Deep Task — see the RUSH guide.