OPEN MODEL INFERENCE

Open models.
One API.

Your next idea deserves a direct path to inference. Build with open language models through an API you already know.

Full privacy. Your queries stay yours.

We never use your API queries to train models.

Start building Read the docs

Google or GitHub signup Access by approval

OpenAI-compatible API One endpoint Open language models Your existing SDK

LESS SETUP. MORE BUILDING.

A familiar API.
A fresh starting point.

Keep your tools and your workflow. Point your OpenAI-compatible client at Dheex and choose your model.

curl --fail-with-body https://api.dheex.com/v1/responses \
  -H "Authorization: Bearer $DHEEX_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-ai/DeepSeek-V4.1-Flash-1017",
    "input": "Hello, dheeX.",
    "store": false
  }'

Make yourself at home

Create your API key

Once approved, generate a key and choose the models it can access.

Build something good

Try a prompt in the playground, connect your app, and follow your token usage.

CHOOSE YOUR STARTING POINT

Open models, ready to build.

Use a stable model name in your requests. Your dashboard shows the models enabled for your account.

LANGUAGE MODEL

deepseek-ai/DeepSeek-V4.1-Flash-1017

1M context

Available through the Dheex Responses API, with streaming responses and token usage reporting.

Request access

Your keys. Your visibility.

Choose model permissions for each key. See requests and input/output tokens in one clear dashboard.

Explore the API

FROM FIRST PROMPT TO WHAT’S NEXT

What will you build?

Get your API key. Find your model. Make it happen.

Request access Beta access is reviewed individually.

Open models.One API.

A familiar API.A fresh starting point.