Start here
Radium documentation, organised by what you're trying to do. Call the API directly, point an existing tool at it, or wire up tracing and evals.

Start here
Radium is an API. There's no chat interface to log into, you connect your existing client, agent framework, or IDE to Radium the same way you'd connect it to OpenAI or Anthropic, and everything downstream keeps working. This page is the fastest way into whichever part of that you need.
Radium API
How to call Radium directly: authentication, models, environments, tool calling, and the errors you'll actually hit.
API quickstart
Tool calling and MCP
Coding agents & IDEs
Point your editor or CLI at Radium.
Run Claude Code on Radium
Use Radium with Cursor
Frameworks & SDKs
Radium inside the frameworks your application already runs on.
Use the OpenAI Python SDK with Radium
Use LangChain with Radium
Use LangGraph with Radium
Use LlamaIndex with Radium
Use CrewAI with Radium
Use DSPy with Radium
Use AG2 with Radium
Use Agno with Radium
Use PydanticAI with Radium
Use the Vercel AI SDK with Radium
Observability, evals & guardrails
Trace it, test it, and guard it the same way you would any other model in production.
Observability and evals with Radium
Everything, by shelf
Radium API — the API itself
Integrations — coding agents, frameworks and SDKs, observability