The LLM gateway is one API endpoint for all of your AI models. It runs on the Anchor Node. It holds every provider key on the server. It gives each person a different key. The software is LiteLLM. The gateway started on 20 August 2026.

​
Why it exists

Open Design keeps its model API key in the browser. It does not keep the key on the server. This causes three problems.
You type the key again in each browser. The gateway cannot change this. Open Design reads the key from browser storage.
Everyone who signs in shares what the browser holds. The gateway solves this. Each person gets a different key.
Without the gateway you add each provider one time in each browser. The gateway solves this. You add one provider and reach every model.

​
What you get

One entry, all models

Add one provider in Open Design. Reach every model behind the gateway.

Your own key

Cancel one person’s key. The other keys continue to work.

Real keys stay on the server

A colleague never sees your DeepSeek key or your MiniMax key.

A cost limit for each person

The gateway stops a key when the person reaches the limit.

​
The two addresses

The gateway has two addresses. They are not the same.
Give api.hjc925.cloud to other people. Do not give llm.hjc925.cloud to other people. The llm address serves the admin pages and the management API.
The api address refuses all admin requests. A test on 20 August 2026 gave these results.

​
Speed

The gateway adds a small delay. The measurement on 20 August 2026 shows this. The delay is 157 ms. This is the second network hop. Your computer sends the request to the Anchor Node. The Anchor Node sends the request to the provider.
LiteLLM itself adds almost nothing. The same test from the Anchor Node shows a delay of 1 ms. The delay that you see is distance, not software.