Does Luz AIbox run LLMs locally?
No. AIbox does not run large language models on-device. It runs the agent orchestration layer (task planning, memory, tool use) locally, while LLM inference happens via cloud API. The reason: running an LLM locally demands an expensive GPU and a lot of power, and cloud models improve faster. Keeping inference in the cloud and orchestration plus memory on-device is the best balance of cost and privacy.
Which cloud LLMs does Luz AIbox support?
OpenAI GPT, Anthropic Claude, Google Gemini, DeepSeek, Qwen, and any OpenAI-compatible API. You can pick the best model per task — a cheaper model for everyday Q&A, a stronger one for complex work.
How much does Luz AIbox cost?
The Pro model is $368 USD as a one-time hardware purchase. The Enterprise edition is custom-priced based on volume and deployment. Credit plans are sold separately (monthly from $5 USD). Across the full eight-model line, prices range from $226 to $1,000 USD — see the product pages for details.
Do I need a subscription to use Luz AIbox?
No. The AIbox hardware itself does not require a subscription and is a one-time purchase. It comes with a small amount of credits out of the box — ready to use, no cloud API key needed. When credits run out, buy a credit plan (monthly, expires end of month, or one-time top-up valid for 365 days).
How are credits charged and consumed?
Credits are consumed based on token usage when your agent makes cloud LLM calls; rates differ per model. Monthly plans reset at month-end, and one-time top-ups stay valid for 365 days. You can monitor usage in your account dashboard.
How do I top up credits?
Three steps: (1) log in or register at the top right; (2) open the credits page, pick a plan, and click Buy; (3) choose a payment method (credit card or PayPal) and complete payment. Credits arrive instantly and are visible in your account.
Is Luz AIbox suitable for enterprise use?
Yes. Typical deployments include 24/7 customer-service agents, office automation, edge AI gateways, and R&D prototyping. The Enterprise edition offers bulk purchasing, private-cloud deployment, and SLA support — contact our sales team directly.
How do I set up AIbox, and how long does it take?
Plug in power, connect to Wi-Fi or Ethernet, and the device auto-activates and pairs — usually chatting within five minutes. It is ready out of the box with no API key configuration needed.
Which chat channels can I use with AIbox?
WeChat, WeCom, WhatsApp, Telegram, and other popular messaging platforms. That means you can talk to the agent inside the messaging app you already use, instead of installing a separate one.
How is my data handled? Is it secure?
Your API keys and local memory stay under your control. No account is required to run, and data never leaves the box without your permission. Memory is stored on the local device rather than a third-party server — the core structural difference between a local agent runtime and a pure cloud solution.
Can AIbox work offline?
It requires an internet connection to function. The agent runtime (task orchestration, memory storage) runs locally, but intelligent conversation and task processing rely on cloud LLMs. Without network access, AI inference is not available.
How much power does AIbox consume?
Typical consumption is under 10W — roughly $8 to $12 USD a year in electricity, less than a standard LED bulb. It is designed for 24/7 always-on operation.
What does the warranty cover, and for how long?
A 1-year (365-day) full hardware warranty covering the motherboard, power supply, and enclosure. Lifetime free OTA updates for system, models, and expert mode. Software and credit plans are not covered by the hardware warranty.
How long does shipping take, and do you ship worldwide?
In-stock items typically ship within 1 to 3 business days; pre-order or custom items list their estimated ship date on the product page. You receive a tracking number by email once your order ships. We ship worldwide to over 100 countries and regions; shipping is calculated by destination and weight and shown at checkout, and some countries may impose customs duties or import taxes that are the buyer's responsibility.
What is your return policy?
A 7-day no-questions-asked return policy. Products must be in original condition with complete packaging and not activated; custom products are not eligible. The buyer pays return shipping, except for defective items.
What payment methods do you accept?
Credit cards (Visa, Mastercard, American Express), PayPal, Apple Pay, Google Pay, and bank transfer. For enterprise bulk orders, contact our sales team to arrange payment terms.
What languages do the website and products support?
The website and interface support 9 languages: Chinese, English, Malay, Russian, Japanese, Korean, Spanish, Portuguese, and French. The device itself works globally with any cloud LLM that supports your language.
What is the difference between AIbox and a Raspberry Pi?
AIbox is a finished, plug-and-play device purpose-built to run an AI agent; a Raspberry Pi is a general-purpose development board you must flash, configure, and debug yourself. If you want "plug it in and go," choose AIbox; if you want "build and modify everything myself," choose a Raspberry Pi.