OwnAI Compute
Book a call Sign in
Local LLM & personal AI experience

Software packages, installed and ready

Optional preconfigured stacks on top of NVIDIA JetPack. Priced separately from hardware. Supported models, context lengths and speed depend on the Jetson platform, memory, quantisation and software compatibility.

Personal AI Package

4 GB+

Everything needed to chat with a local model from any browser in the house.

$149one-off
  • Local LLM runtime (Ollama / llama.cpp)
  • Open WebUI chat interface
  • Model management & quantisation presets
  • Basic document upload

Private Knowledge Package

8 GB+

Turn your documents into a searchable, question-answering knowledge base.

$249one-off
  • Local LLM
  • Local embeddings
  • Vector database (Qdrant or Chroma)
  • Document indexing pipeline
  • RAG chat interface
  • Searchable knowledge base

Voice AI Package

8 GB+

A private voice assistant with no cloud round-trips.

$199one-off
  • Speech-to-text (Whisper / Riva where supported)
  • Text-to-speech (Piper / Riva)
  • Voice interaction interface
  • Wake-word support
  • Optional smart home integration

Developer AI Package

4 GB+

A ready-to-code AI lab with containers, APIs and monitoring.

$179one-off
  • Docker + NVIDIA Container Toolkit
  • Model deployment tools
  • OpenAI-compatible local API endpoints
  • JupyterLab development environment
  • Monitoring dashboard (Prometheus + Grafana)

Smart Home AI Package

8 GB+

Natural-language control of your home, processed locally.

$229one-off
  • Home Assistant integration
  • MQTT broker
  • Compatible device integrations
  • Local automation engine
  • AI-powered command interface

Complete Personal AI Package

16 GB+

The full stack: assistant, knowledge base, voice and smart home in one managed install.

$499one-off
  • Personal AI assistant
  • Local knowledge base
  • Voice interface
  • Smart home integrations
  • Model management
  • Optional secure remote access

Compatible ecosystems

  • NVIDIA JetPack
  • CUDA (where supported)
  • TensorRT
  • TensorRT-LLM (where supported)
  • Ollama
  • llama.cpp
  • Hugging Face tooling
  • Docker + NVIDIA Container Toolkit
  • Open WebUI
  • Qdrant / Chroma
  • Whisper
  • Piper / Riva TTS
  • Home Assistant
  • MQTT

What we promise, and what we don't

Your models, documents and voice stay on your hardware for supported workloads. We do not promise that cloud-scale models run locally: a 4 GB station runs ~3B models, 8 GB runs 7–8B at INT4, 16 GB reaches 13B, and only 64 GB+ modules can hold 30B-class models. Package minimum-memory badges reflect that.

Rejoining the server...

Rejoin failed... trying again in seconds.

Failed to rejoin.
Please retry or reload the page.

The session has been paused by the server.

Failed to resume the session.
Please retry or reload the page.