FAQ

Frequently asked questions organized by topic. If you can't find an answer, search Getting Started and the feature guides first.

Installation & Login

Which operating systems are supported?

Windows 10+ (x64), macOS 12+ (Apple Silicon / Intel), Ubuntu 20.04+, with RAM ≥ 8GB. See Installation & Download for download and setup steps.

Why am I suddenly asked to solve a captcha at login?

The captcha is trigger-based and does not appear normally: it is required only after 2 consecutive login failures, or when too many requests in a short time trigger rate limiting (429) — a protection against brute force. Pass it by clicking the positions in the order given by the text in the image.

How long does the login session last? Is verification-code login supported?

The login token is managed centrally by the local kernel and renewed automatically — no re-login needed after restarting the IDE, and no extra "7-day auto-login" switch. The "Verification Code Login" tab on the login page is currently greyed out (coming soon); please log in with email + password — see Register & Login.

Models & Billing

Which models are supported?

15+ mainstream models including Kimi K3, GLM-5.2, DeepSeek-V4, Qwen3.6, Doubao and MiniMax, via official direct and Volcano Engine channels — see "Settings → Models" for the definitive list. Details: Supported Models.

How are Agent runs counted? Does chat consume them?

Only Agent runs count against the quota: each RUSH task counts as 1 run regardless of internal iterations; AI chat, inline edit and FIM completion never consume runs — use them freely for everyday coding. Details: Plans & Quota.

When does the daily quota reset? Do unused runs roll over?

The daily quota resets automatically at 00:00 Beijing time (UTC+8); unused runs do not roll over to the next day. Today's usage and the reset time are visible in real time in "Settings → Subscription & Usage".

Can I bring my own API key? Is it safe?

Yes. After binding a vendor key in Settings, requests are billed at the vendor's official prices with zero platform markup; keys are AES-256 encrypted at rest and protected by 2-of-3 key sharding — never stored in plaintext. Details: BYOK.

Is there a yearly-billing discount?

A yearly plan costs about 10 months of monthly billing — Pro ¥290/year, Premium ¥990/year, saving ~17% — and takes effect immediately after upgrading. See the pricing page.

Features & Shortcuts

What are the shortcuts for RUSH and inline edit?

RUSH deep task: Ctrl + Shift + R; inline edit: select code and press Ctrl + K. Full list: Keyboard Shortcuts.

Can I migrate from VS Code / Cursor?

Yes. The first-launch migration wizard imports keyboard shortcuts, themes, settings and common plugin configurations in one click — up and running in 5 minutes. Details: Migrate from VS Code.

Security & Kernel

Is my code uploaded to the server?

No. LingXizhi is local-first: your code stays on your device by default; only the necessary context is encrypted and sent to the model you choose when you make an AI request. Sensitive files such as .env and credentials.json are excluded by default (see privacy.exclude_files in Configuration), and sensitive data such as API keys is AES-256 encrypted at rest.

What if kernel port 3721 is occupied?

No manual handling needed: the kernel listens on 127.0.0.1:3721 by default; if the port is occupied it auto-increments (up to 10 attempts) and writes the actual port back to configuration. Run curl http://127.0.0.1:3721/health to check kernel status; main endpoints: Kernel API.

Next Steps

• Still unresolved? Check your local setup in Configuration, then get the latest version from the download page and reinstall.
• For feature details, continue with the feature guides.