Local AI workstation for private inference with up to 192 GB of unified memory and 32 GB Radeon GPU
Infrastructure for real-time AI agents
Stage
Not given
Established
Founded
Not given
2009
Based in
Not given
US
Pricing model
Paid
Paid
Pricing
Developer: starting at $5,999 USD (128 GB memory, 2 TB storage). Memory upgrade to 192 GB adds $2,200. Storage upgrade to 4 TB adds $250. Cluster kit $799 per machine. Business pricing available by quote. Worldwide shipping, import duties and taxes outside US included. No sales tax in US.
Voice AI Agents: from $0.05 per minute. Text-to-Speech: from $0.000003 per character. Speech-to-Text: from $0.0015 per minute. Voice API: from $0.002 per minute. SIP Trunking: from $0.0032 per minute. SMS and MMS: from $0.004 per message. WhatsApp: from $0.0025 per minute. Email API: from $0.30 per
Key features
Run DeepSeek V4 Flash and Qwen 27B models
Browser-based setup over encrypted Bluetooth
Private OpenAI-compatible API
Lucebox Mix: run two models simultaneously
Headless Ubuntu Server with pre-installed inference engine
Optional Tailscale integration for remote access
Optional cluster kit with 100 GbE networking
Voice AI agents with speech-to-text, text-to-speech, and inference on one control plane
Serverless and dedicated GPU inference with 50+ open-source models
Edge compute with Functions, Stateful Actors, KV, SQLDB, and Object Storage at 20+ points of presence
Global phone numbers, toll-free numbers, and number porting
Owned carrier infrastructure with 25+ points of presence and 30+ carrier licenses
OpenAI SDK compatible inference API with streaming and tool calling
Agent SDK for building stateful AI agents with memory and scheduled tasks
What makes it different
No per-token billing or cloud dependencies
1.5x faster decode speed than DGX Spark
Hardware and software co-designed for optimal performance
Completely private: runs offline with no cloud proxy
Open source inference engine with tuned models
Only licensed carrier that owns AI infrastructure end-to-end including private network, GPUs, and edge compute
Everything on one control plane and one bill without multiple vendors
Inference runs colocated with media plane on owned GPUs with no egress fees