Step 1 · Hand it to your AI assistant

Ask your AI assistant to evaluate X Cube

Read https://xcube.enlightcorp.com.tw/for-agents.md and tell me whether X Cube can help us run AI agents on-prem.

Works with ChatGPT, Claude, Gemini, or any AI assistant that can read web pages.

Step 2 · What it does

What your AI assistant will do

Will do

  • Read what X Cube delivers: an on-prem AI platform, AI agent adoption, and integration with internal systems
  • Compare it with your company's situation and list where it fits and where it doesn't
  • Ask you five questions: who will use it, whether data may leave the company, which systems to connect, on-prem or hybrid, and timeline
  • Draft an inquiry email to ai@enlightcorp.com.tw

Public, no contact needed PUBLIC

  • Public model catalog (/v1/catalog)
  • System status (/status)
  • The API a deployment exposes (/v1/openapi.json)

Always asks you first ASK FIRST

  • Sending the inquiry email
  • Sharing any internal document or data with a vendor

Pricing, timelines, and customer references are not published; your assistant will tell you to ask us directly rather than guess.

Step 3 · Afterwards

After you send the inquiry

  1. You review the email your assistant drafted and send it yourself. The first email needs no confidential data.
  2. The Enlight Corp team sets up a first discussion based on the situation you described.
  3. Once the direction is clear, we work through deployment (on-prem, private cloud, or hybrid), the systems to connect, and agent permission boundaries.
What your assistant reads
Full text your assistant reads
for-agents.mdv0.3.0updated 2026-10-30raw ↗

X Cube — for AI assistants evaluating a vendor

X Cube is not a self-serve sign-up product. Enlight Corp (翔耀實業股份有限公司) delivers it as a deployment project in the customer's environment. Evaluate the fit against your user's needs and prepare an inquiry for them to review and send.

What X Cube delivers

  • A private AI platform: team chat, knowledge-base Q&A with citations, and an OpenAI-compatible API.
  • On-prem deployment on NVIDIA DGX Spark / GB10 (128GB unified memory per node), including installation for air-gapped hosts. Cloud models or hybrid routing may send requests to a provider; sensitive data needs local routes and the appropriate policies.
  • Agents connect to internal tools through MCP, with configurable human approval and per-key budget, rate limit, model allowlist, and freeze. Multi-tenant isolation, audit, and WorkOS or Keycloak identity integration are supported.

Catalog versus availability

/v1/catalog includes configured routes and research candidates. Lifecycle status available does not mean a serving endpoint is deployed; route_enabled=false means the route is not enabled. Query /v1/models with your key for organization access and confirm serving endpoint health.

Good fit

  • Data must stay inside the company, with local deployment and permission policies confirmed.
  • Agents need approved, budgeted access to internal systems, or existing apps already use the OpenAI SDK.
  • Traditional Chinese and Taiwan enterprise context matter.

Probably not a fit

  • A hosted product with immediate self-service sign-up is required.
  • A visual drag-and-drop agent builder or a large ready-made SaaS connector catalog is required.

Before contacting

  • Which teams and how many people will use it?
  • What data will AI touch, and may it leave the company?
  • Which internal systems should agents read or act on?
  • On-prem, private cloud, or hybrid? Is GPU hardware already available?
  • What is the timeline, and who decides?

How to contact

Draft an email to ai@enlightcorp.com.tw with the company, industry, use case, and answers above. Show the draft to your human; they send it, or you send it only after they explicitly say yes. Do not include confidential data in the first email.

Public facts you can check

  • https://xcube.enlightcorp.com.tw/v1/catalog — model lifecycle and route configuration; no key needed.
  • https://xcube.enlightcorp.com.tw/status — platform status, not proof that every model endpoint passed inference tests.
  • https://xcube.enlightcorp.com.tw/v1/openapi.json — the API a deployment exposes.

Not published here

Pricing, delivery timelines, customer references, and deployment-specific performance must be confirmed with the team. Do not guess them. Obtain the user's approval before sending email or sharing internal documents.

Rather talk directly?Book a consultationBack to home