HEANEXUS
AI Work Environment · Local-first

The AI Work Environment,
on your own compute.

HEANEXUS is packing a governable AI work environment onto a USB device: plug it into your own machine or server and run your own models on your own compute. Your workspace and configuration stay on devices you control; when you deliberately choose a cloud model provider, the prompts and context required for that call are transmitted to that provider — you choose the boundary. The USB stick is the delivery medium, not the inference hardware.

Stage: Engineering validation (EVT) — not commercially released
Development status Partnerships

What it is

HEANEXUS follows a DSH-first Hybrid strategy: we build on the open-source DeepSeek Harness as the runtime foundation, and focus on productization, delivery, and enterprise governance — we do not rewrite a general-purpose agent core.

Local-first, your boundary

Your workspace, configuration, and files remain under your control. With local models, inference runs on your devices; when you deliberately connect a cloud model provider, the required context is transmitted to that provider.

USB delivery (EVT)

A USB device carries the runtime environment and license, aiming at deploy-and-migrate by plugging in. Currently an engineering sample — hardware validation is not yet complete.

Governance by design

Policy, evidence, and audit are designed into the architecture rather than bolted on — built for enterprise adoption, delivered milestone by milestone, not yet implemented in production.

Bring your own compute

HEANEXUS is not tied to any cloud and does not bundle compute sales. Three typical sources:

Local GPU workstation

e.g. a self-built dual RTX 3090 host running open models — one typical form of local inference compute (your own environment, not a shipped HEANEXUS product capability).

Your own cloud hosts

Cloud servers you already own can act as remote inference nodes, under your own network and access controls.

API providers

Connect third-party model APIs with your own keys; with a cloud provider, the required context is transmitted to that provider. Production-grade secure persistent key storage is still in engineering validation.

Current product direction

Four product forms advance along the roadmap, all in engineering validation or design. Detailed status and validation evidence live on the status page.

HEANEXUS Key

USB delivery medium + local runtime + license verification. Currently an EVT engineering sample; hardware validation (T-001B) has not started. EVT

HEANEXUS Local

Local model integration, local knowledge base, offline capable. Early engineering validation; Alpha validation not yet complete. EVT

HEANEXUS Anywhere

Managed remote access to your own Host (phone/tablet/desktop). In design, not available. PLANNED

Enterprise governance

Ontology, RBAC, policy, and audit domains are designed into the architecture and delivered by milestone — not yet implemented in production. PLANNED

Note: All capabilities described on this site are in engineering validation or design. Nothing here constitutes a shipping commitment. Final SKUs and pricing have not been announced.

Contact & partnerships

We are talking to partners in: model & inference infrastructure, enterprise on-premise AI deployment, and channel/delivery cooperation.

Contact channels being provisioned: official mailboxes ([email protected] / [email protected]) are NOT yet operational — please do not send sensitive information to these addresses. This page will switch to a working channel once email service is live. In the meantime, please reach us through existing business channels.