AIAionRay
Log inGet started
AI service operations platform

Turn your GPUs into
callable AI services.

AionRay helps teams connect owned GPUs and idle compute, then run training, deployment, OpenAI-compatible APIs, app monetization, compute rental, billing records, revenue records, and troubleshooting from one console.

Get startedTry playground
OpenAI-compatible API·Model fine-tuning / custom training·Usage and revenue records·Private assets stay in your controlled environment·One-command GPU onboarding
Capabilities

Everything you need to ship an AI product, in one chain.

Training, API, billing, and revenue records should not live in separate tools. Put them together and you ship faster.

Train → Deploy
From data to callable API
  • Upload data or point to a local dataset
  • Launch fine-tune jobs on attached machines
  • Publish as an inference service in one click
API → Billing
Request-ready endpoints, usage-based billing
  • Works with common OpenAI client calls
  • Per-request or per-token billing pipeline built in
  • End-user credentials, quotas, and bills centralized
App → Revenue
Turn your API into a sellable product
  • Wrap the endpoint as a published app
  • Configure access policy, price, quota
  • Subscription and usage revenue are tracked in the ledger
Common API support

OpenAI-style clients
connect with simple config.

AionRay implements the Chat / Completions / Embeddings / Models routes. Swap the base_url and key, then keep your existing SDK flow for core requests.

/v1/chat/completions
/v1/completions
/v1/embeddings
/v1/models
Compatibility scope

The public contract covers the core `/v1/*` routes below. Advanced request fields are forwarded to the underlying model when the selected service supports them.

Canonical public routes: `/v1/chat/completions`, `/v1/completions`, `/v1/embeddings`, `/v1/models`.Tool calls, structured output, and multimodal content depend on the model and runtime you pick.
client.py
from openai import OpenAI

client = OpenAI(
    base_url="https://your-gateway/v1",
    api_key="sk-xxxxxxxx",
)

resp = client.chat.completions.create(
    model="your-model",
    messages=[
        {"role": "user", "content": "Hello"},
    ],
)
One platform vs. stitching tools

One console for training, deployment, billing, and distribution.

Keep the core AI product workflow in one place, instead of switching between training, serving, metering, and revenue accounting tools.

What you need to do
Typical approach
AionRay
Fine-tune and deploy as API
Train on one platform, deploy on another
Same console
Per-token or per-request billing
Build billing and usage tracking yourself
Built-in per-token billing pipeline
Publish API to third parties
Build auth and revenue accounting yourself
Built-in app credentials and revenue records
Usage and revenue reconciliation
Manually stitch usage, bills, and revenue records
Ledger records and audit-ready usage flows
Who it's for

Who uses AionRay

Developers shipping an AI app

Turn an idea into a subscription AI tool — from training to launch, no tool switching.

Teams building domain models

Ship a domain model as a metered API. Access policy, quota, audit, and billing are ready out of the box.

Internal AI service teams

Package internal models into services for internal or external use, with quota, billing, and audit in one place.

5 steps

5 steps to get going

  1. 1
    Sign up and open the console
  2. 2
    Connect your own GPU, or rent one from the marketplace
  3. 3
    Upload data, start a fine-tune
  4. 4
    Deploy the inference service, or publish as an app
  5. 5
    Set pricing and budgets, then start billing

Start from your own GPU or model.

Connect compute, train or import a model, deploy it as an API, package it into a subscribable app, and keep operations visible from one console.

Get startedTry playground