Skip to main content

How do I use the Wiro LLM Gateway (Completions API) in VS Code, Cursor and other AI tools?

Connect VS Code, Cline, Cursor, Claude Code and other OpenAI- or Anthropic-compatible tools to Wiro's chat models with one API key.

Wiro's Direct LLM Gateway gives you 50+ chat models from the GPT, Claude, Gemini, Grok, DeepSeek, Qwen, Llama, Mistral, GLM and Seed families, including uncensored models, through one OpenAI-compatible API, with Anthropic Messages support for Claude Code. You can use them in VS Code's built-in chat (no extension needed), in any extension that accepts a custom OpenAI base URL such as Cline, Roo Code or Continue, and in Cursor, Claude Code and other coding tools. See wiro.ai/llm for more.

Easiest setup: follow the Connect an agent section of our docs. It has step-by-step guides for Cursor, Claude Code, VS Code, Windsurf, OpenClaw, Hermes and OpenAI SDK.

What you need

  • API key: create a project in your Wiro dashboard (see How do I get an API key?). Both project types work, but the credential differs: API Key Only uses the key alone, Signature-Based uses YOUR_API_KEY:YOUR_API_SECRET (the key alone is rejected). For a project used only with coding tools, API Key Only is the simplest.

  • Base URL: https://llm.wiro.ai/v1 (Anthropic-style clients such as Claude Code use https://llm.wiro.ai).

  • Auth header: Authorization: Bearer YOUR_API_KEY (for a Signature-Based project, Bearer YOUR_API_KEY:YOUR_API_SECRET). Most tools send this for you from their API key field.

  • Model ID: lowercase owner/model, for example openai/gpt-5-6-sol, claude/sonnet-5 or qwen/qwen3-8-27b-uncensored. Use the ID exactly as it appears in the catalog or in /v1/models. The Run API name shown on a model's page (for example Qwen/Qwen3.8-27B-Uncensored) is a different format and won't work here.

To list every model ID your project can use:

curl https://llm.wiro.ai/v1/models \
  -H "Authorization: Bearer YOUR_API_KEY"

You can also browse them in the model catalog.

Set up your tool

  1. Choose the OpenAI Compatible (or custom endpoint) provider.

  2. Enter the base URL https://llm.wiro.ai/v1. If the tool asks for the full endpoint, use https://llm.wiro.ai/v1/chat/completions.

  3. Paste your credential into the API key field.

  4. Add the model ID you want to use, for example openai/gpt-5-6-sol.

  5. If the tool asks for limits, use the model's context window and max output from /v1/models.

VS Code built-in chat: open Chat: Manage Language Models, then + Add Models, then Custom Endpoint. Set the API type to Chat Completions and use the full endpoint URL. Cline: pick "OpenAI Compatible" (not "OpenAI") and tick "Computer Use" to enable tool calling.

Quick test

curl https://llm.wiro.ai/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "openai/gpt-5-6-sol", "messages": [{"role": "user", "content": "Hello"}]}'

Good to know

  • Billing: calls are billed to your Wiro account through the same billing system as any other run. When usage data is available, the response also reports the charged amount in usage.cost.

  • Tool calling: standard OpenAI function tools work. Your editor or agent runs the tools, not Wiro.

  • Streaming: "stream": true is supported, but in the current release the streamed events are sent after the model has finished its reply, so text won't appear token by token.

  • Uncensored content: see Which models support uncensored or NSFW content? for details.

  • Never put your key in browser or mobile app code.

Related Articles
Did this answer your question?