InferencePreview

Hosted models behind an OpenAI-compatible API.

EZGH Inference serves models behind an OpenAI-compatible API. The stock OpenAI SDKs work with only the base URL and API key changed.

Open the console Documentation

Inference is in preview while model capacity comes online.

embed.py

from openai import OpenAIclient = OpenAI(    base_url="https://inference-api.ezghcloud.com/v1/…/openai/v1",    api_key=os.environ["EZGH_API_KEY"],)r = client.embeddings.create(    model="qwen3-embedding-0.6b",    input=["what is a vector database"],)

What you get

OpenAI-compatible

Use the OpenAI SDKs for Python and TypeScript, or plain HTTP. Set the base URL and your EZGH API key; nothing else changes.

Multilingual embeddings

qwen3-embedding-0.6b: 1,024 dimensions (down to 32), 8,192 tokens per input and 128 inputs per request.

A public Model Catalog

Every model, its versions and limits, readable without a credential. Pin a version such as :1.0.0 or follow the latest.

Nothing kept

Inputs and embeddings are never stored or logged.

Pricing

All pricing

Input tokens

$0at launch

Usage is metered by input token.

At a glance

Requests: 600 a minute per project
Input tokens: 1,000,000 a minute per organization
Monthly allowance: 100,000,000 input tokens
API: inference-api.ezghcloud.com

Built on the platform

Governed like every other product, from the first call.

IAM policies Trails audit log One monthly bill Service quotas