Use SinarCloud in VS Code

Several coding agents for VS Code can run on SinarCloud models. Each one takes the base URL and your API key in its own settings, and none of them needs an account with the extension’s maker. Pick the one you prefer, then follow its guide. Claude Code uses the Anthropic Messages format. The other three use the OpenAI Chat Completions format. Both formats use the same API key, models, and balance. See Compatibility.

​

Before you start

• Get an API key and a model alias as described in Before you connect.
• Create a separate API key for each extension. The Usage page can then show what each one spends, and you can revoke one key without breaking the others.
• Choose a model whose Model capabilities list Chat completions, Tool use, and Streaming.

​

What all of them have in common

• Each step is a request. An agent sends a new request for every step of a task, and each one includes the conversation so far. Start a new task or session for each new piece of work.
• Their cost displays aren’t your charges. The extensions estimate costs in their own way, or show only tokens. To see what you spent, open AI > Usage in the console and filter by the extension’s API key.
• Output limits must fit the model. If you set a maximum output in the extension, keep it at or below the model’s Maximum output (tokens). Otherwise SinarCloud refuses the request with “The output-token limit exceeds the model’s allowed range.”

​

Guides