Hubrium
Sign in

Developer quickstart

From key to first token in three steps.

Keep your existing OpenAI or Anthropic SDK and change only the endpoint and API key.

01

Quickstart

Pick the client you already use. Every tab uses the same Hubrium key, base URL and model IDs.

1

Create an API key

Create a scoped token in the Hubrium console. Treat it like a password and keep it server-side.

Get API key
2

Configure your client

Pass the Hubrium endpoint as base_url to your OpenAI or Anthropic client.

3

Send a request

Choose any enabled model ID from the models page or models endpoint.

curl https://api.hubrium.com/v1/chat/completions \
  -H "Authorization: Bearer $HUBRIUM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v3.1",
    "messages": [
      {"role": "user", "content": "Explain this code."}
    ],
    "stream": true
  }'
02

Core endpoints

OpenAI-compatible Chat Completions and Responses alongside Anthropic-compatible Messages. Model capabilities depend on the current gateway configuration.

MethodPathPurpose
GET/v1/modelsList models available to your API key
POST/v1/chat/completionsOpenAI chat, tools and streaming
POST/v1/responsesOpenAI Responses-compatible agent requests
POST/v1/messagesAnthropic Messages, tools and streaming
03

Made legible to agents

Hubrium publishes stable, machine-readable discovery files so coding agents can understand models, prices and the API without scraping the interface.

/llms.txt

A concise map of product and documentation content.

/models.json

Current model IDs, context windows and per-token pricing.

/openapi.json

An OpenAPI description of the compatible API surface.

Need implementation help?

Send us your SDK, expected volume and target models. We will help map the cleanest migration.

[email protected]