HYBT / SYNQAI OS / ON KICKSTARTER

Coming soon to Kickstarter

SYNQAI OS

Your Jetson. Your private AI tutor.

A prepared microSD card for your NVIDIA Jetson Orin Nano 8GB Developer Kit. JetPack, a local 4B language model and the SYNQAI tutor interface, brought together so you can start learning instead of assembling the software stack.

Explore the prototype
Downloads and the local terminal
The prototype tutor interface. An original phone photograph, framed and straightened for this page. Open full photo

01 / THE VALUE

The engineering work, already brought together.

The value is the preparation and integration: an AI learning environment on hardware you own, with a path from connected setup to local use.

A prepared software stack

JetPack, a local 4B language model and the tutor UI are integrated for the Jetson Orin Nano 8GB Developer Kit.

Network controls within reach

Wi-Fi setup controls are part of the prototype. Connect to your own Wi-Fi or Ethernet network when you choose to download materials or use online services.

Your own learning library

Use locally available course materials. Add repositories, tutorials and compatible models through the terminal when connected, then keep the downloaded resources on your device.

02 / THE LOCAL TERMINAL

Download your learning library. Then go local.

SYNQAI OS includes access to a local terminal, not just a chat window. Use it to prepare your own library: course repositories, tutorial files and compatible model weights.

  • Clone a course repository while connected and keep its downloaded files on the Jetson.
  • Download tutorials and supporting files you are licensed to use. External videos and links may need separate downloads or a connection.
  • Add compatible models with the appropriate runtime. Check the model licence, available storage and memory first.
See the terminal photograph

ILLUSTRATIVE DOWNLOAD COMMANDS / NOT PHOTOGRAPHED OUTPUT

mkdir -p ~/Courses
git clone --depth 1 https://github.com/microsoft/generative-ai-for-beginners.git ~/Courses/generative-ai
ls ~/Courses/generative-ai

Example: downloading a public course repository. Requires Git and an internet connection. Downloaded course files can be kept locally; exercises that call cloud services still need their own connection and setup. This example is not a claim that every exercise runs on the supplied image.

Example course repository and licence

03 / ALL NINE PHOTOGRAPHS

The complete prototype workflow, in photographs.

All nine photographs in capture order: startup, tutor, network controls, a question and response, console access, terminal, and shutdown. Open any photograph to inspect the full publication copy.

Photographs remain in their capture order. Perspective is corrected for display. Privacy blocks are baked into the network and console publication files; their unredacted versions are not linked or shipped with this page. Pixels outside the privacy blocks are unchanged. No commands or download progress have been added to the photographs.

04 / YOUR LEARNING LIBRARY

Connect to prepare. Disconnect to learn.

Local work starts with local resources. Download what you need before leaving the network.

  1. Boot the prepared card

    Use your compatible NVIDIA Jetson Orin Nano 8GB Developer Kit. The Developer Kit is not included.

  2. Choose your materials

    When connected, download the repositories, tutorials, course files or compatible models you need. Check their licences, storage and memory requirements.

  3. Set your network boundary

    For offline use, disconnect Ethernet and disable Wi-Fi and other network connections. Cloud models, live search and new downloads require a connection.

  4. Study locally

    Use the preloaded local model and materials already saved on the device. You decide when to reconnect.

05 / YOUR CONTROL

Local first. Clear about the boundary.

Sovereign AI, private AI and an air gap describe different things. Here is what they mean for this project.

SOVEREIGN AI

Control the local setup

Choose the materials, compatible local models and operating settings on your own Jetson. Control comes from how the system is configured and operated.

PRIVATE AI

Keep the local path local

Local mode is designed to use the model on your device without sending prompts to a cloud model. If you enable an optional cloud model, data goes to the provider you select.

AIR GAP

Isolation is an operating choice

Offline mode alone does not certify an air gap. An isolated deployment requires disconnected network links, disabled radios and controlled transfers of files or removable media. Online downloads and cloud services sit outside that boundary.

What “air gap” means — NIST glossary

For learning that needs local control.

Local course materials can matter in education, public-sector training and health-related staff learning. Each organisation must assess its own data, security and sector requirements. This project does not claim regulatory certification or approval for sensitive institutional data.

06 / BEFORE YOU BACK

What to know before backing.

  • This is a consumer product in development. The photographs show a working prototype, not the final release image.
  • Your Developer Kit needs firmware compatible with the supplied JetPack image. First-time firmware preparation may be required on older kits.
  • The 4B model is a compact local assistant. Answers can be wrong; check important information against the original learning material.
  • The Developer Kit and other required hardware are not included. Card options are listed below.
  • Prices and final reward contents will be published with the Kickstarter campaign. No price or cloud-model allowance is announced here.
  • Requires an NVIDIA Jetson Orin Nano 8GB Developer Kit (not included).
  • microSD cards: 64GB, 256GB or 1TB.
  • Estimated delivery: January 2027.

SYNQAI OS: Private AI Tutor for NVIDIA Jetson Orin Nano

Bring your hardware. Keep control of your learning.

Follow SYNQAI OS when the Kickstarter page becomes available.

Kickstarter is not a store: backers support projects in development. Delivery dates are estimates.