Skip to content
PRIVATE BETA Built for your next idea

Many models.
One place
to build.

Explore AI models, try your prompts, and connect your applications. One API, with usage and spending under your control.

Enter private beta Explore models
Invitation only. Administrator-assigned credits.
YOUR WORKSPACESample preview
Playground01 / EXPLORE
YOU

What could we build next?

SAMPLE RESPONSE

A small idea. A working prototype.
Something that makes someone’s day easier.

Let’s start with a prompt.

One API. Your choice of model.
You set the boundaries.Usage visibility. Spending controls.
Illustrative interface · no live request is sent
One APIBrowser PlaygroundSpending controlsYour own API keys
MORE WAYS TO THINK

Find your next model.

Different ideas call for different models.
Explore a selection from our text-model catalog.

TEXT MODEL

DeepSeek V4 Flash

Explore a different approach to your next text task.

ai/deepseek-v4-flash-0731Explore in workspace
TEXT MODEL

Claude Sonnet 4.6

Bring your prompts, ideas, and conversations.

ai/claude-sonnet-4-6Explore in workspace

Catalog selection · Availability and supported features are shown in your workspace.

FROM CURIOSITY TO CODE

Less setup.
More making.

A workspace for the whole journey.
Try an idea. Connect it. See how it’s doing.

01 / EXPERIMENT

Start with a conversation.

Use the Playground to test prompts and explore responses before writing your integration.

Explore the Playground
02 / CONNECT

Your key. Your application.

Create separate API keys for your applications, set an expiry, and revoke access when you need to.

See a request example
03 / UNDERSTAND

Keep your usage in view.

Review tokens, request status, and recorded costs. Set spending limits and export usage records.

Open your workspace
BUILT FOR YOUR WORKFLOW

An API that
feels familiar.

Keep the shape of your integration. Set your base URL, add a VGAIN key, and choose an enabled model.

POST/v1/chat/completions
GET/v1/models
Open the API guide OpenAI-style Chat Completions. Supported parameters vary by model.
curl "$VGAIN_BASE_URL/chat/completions" \
  -H "Authorization: Bearer $VGAIN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "ai/deepseek-v4-flash-0731",
    "messages": [
      {"role": "user", "content": "What could we build next?"}
    ],
    "max_tokens": 128
  }'
Set VGAIN_BASE_URL to your approved API address ending in /v1, and VGAIN_API_KEY to your own key.
YOUR FIRST REQUEST STARTS HERE

A small start.
A world of possibilities.

Private access, a personal workspace,
and a clear path to your first request.

  1. 01

    Accept your invitation

    Register using the email address your administrator invited.

  2. 02

    Receive beta credits

    Your administrator assigns your starting allowance and limits.

  3. 03

    Try a model. Make it yours.

    Explore the Playground, then create an API key for your application.

ROOM TO EXPLORE. BOUNDARIES YOU SET.

Build with
a clearer view.

Access that belongs to you

Personal accounts, revocable API keys, and authenticator protection. MFA is required for administrator controls.

Limits before surprises

Workspace and API-key controls help keep spending, request counts, and token usage within configured limits.

Visibility without a prompt archive

VGAIN records request metadata, not prompt or response bodies. Live inference is processed by external model services.

A FEW THINGS TO KNOW

Before you
start building.

How access, credits, and the API
work during our private beta.

Who can use VGAIN Compute?

The private beta is for invited users on the approved network. Ask your platform administrator for an invitation, then register with the email address it was issued to.

How do beta credits work?

Your administrator assigns credits to your workspace. New accounts start with zero credits. There is no checkout or subscription during the beta, and beta credits have no cash value. Check your workspace for the configured usage rates and limits.

Can I use my existing API integration?

VGAIN provides OpenAI-style /v1/models and /v1/chat/completions endpoints. Set your application’s base URL and VGAIN API key, then choose an available model ID. Supported parameters and capabilities vary by model.

Does every model support streaming?

Streaming is available on supported routes. Check the model’s capabilities in your workspace before enabling it. A model appearing in the catalog does not mean every feature is enabled.

What happens to my prompts?

VGAIN does not retain prompt or response bodies in request history. It records metadata such as the model, token usage, status, and cost. Live inference is processed by external model services and is subject to their data policies. Playground conversations disappear when you leave the page.

Is this a GPU hosting service?

VGAIN Compute gives you access to AI models through an API and a browser Playground. The current beta does not include GPU rentals, model training, or dedicated model hosting.

BRING YOUR NEXT IDEA

A prompt is a good
place to start.

Enter private beta

Already have a workspace? Sign in