Skip to main content

What's the difference between Wiro's MCP server and the LLM Gateway?

MCP lets your AI assistant find and run Wiro models for you, while the LLM Gateway lets your coding tool use a Wiro chat model as its AI.

MCP lets your AI assistant or code editor find and run Wiro models for you, while the LLM Gateway lets your coding tool use a Wiro chat model as its AI.

  • MCP server (https://mcp.wiro.ai/v1): your assistant searches the catalog, checks a model's inputs and price, runs it (images, video, text, audio, 3D and more) and follows its tasks. The server itself is free; you pay standard model prices, and only successful tasks are billed.

  • LLM Gateway (https://llm.wiro.ai/v1): 50+ chat models, including GPT, Claude, Gemini and Qwen, through one OpenAI-compatible API, with Anthropic Messages support for Claude Code. Calls are billed like any other run.

Both use your Wiro project's API key, plus its secret for a Signature-Based project. Hosted MCP and the Gateway take it as Authorization: Bearer; local MCP reads it from environment variables. Setup: our MCP guide and LLM Gateway guide, or the docs for MCP and the LLM Gateway.

Did this answer your question?