What is MindBox?+
MindBox is a local AI LLM machine that integrates most of the available open source LLMs
in the market. A perfect taste of AI that balances data security and efficiency.
Why MindBox?+
Save your effort for setting up your own environment for LLM. MindBox comes with a
user-friendly web-based UI, supports multiple users and provides 2T storage for your own documents uploaded
to the machine that requires LLM to navigate.
What LLMs are supported by MindBox?+
MindBox integrates most of the available open source LLMs in the market including
DeepSeek R1, Gemma 3 from Google, Mistral Small 3, Phi-4 from Microsoft,
Llama 3.2 from Meta and QwQ from Alibaba. You are well set to choose among these models and
switch over to any one of them at your wish.
What are the key benefits of MindBox?+
MindBox is your local LLM machine that operates without connecting the machine to the
Internet. All your documents uploaded or chat history will all be within your 100% control, ensuring your
data security and privacy.
Can MindBox be used for SME and how it works?+
Perfect for SME, MindBox only requires minimal plug and play set up at your organization,
just like a NAS. It supports multiple users and comes with necessary account management capability. You can
also access from your desktop, mobile phone and tablet using common web browsers.
Any future upgrade to newer LLMs?+
The AI world is highly competitive and rapid advancing. From time to time, there will be
new LLM available to the market. We will provide timely free upgrade online within the first year
upon purchase.
What languages can MindBox support?+
Currently English, traditional and simplified Chinese.
What sorts of file format that MindBox supports? Can I upload photos or video to MindBox?+
Currently supports Word files, PDF, TXT and image files [Gemma 3 only], and will support
Excel, PowerPoint in the near future.
Do you have more powerful machine [run larger LLMs] with better processing power?+
Yes. We also provide the rack-mount MindServer that comes with 2 GPU cards.
Basically this is a custom model and will be quoted on case by case basis.