Knit guides
Connect your AI and your tools
Step-by-step guides to connect your own LLM providers and MCP servers to Knit. Bring your own key for zero-credit runs, then describe a circuit and export to KiCad.
LLM provider
Anthropic (Claude) to Knit
Use your own Anthropic API key with Knit to generate KiCad schematics with Claude. Bring-your-own keys cost zero Knit credits.
Read guide →LLM provider
OpenAI (and OpenAI-compatible APIs) to Knit
Use your own OpenAI API key with Knit, or point the OpenAI provider at any OpenAI-compatible endpoint by setting a custom base URL.
Read guide →LLM provider
Google Gemini to Knit
Use your own Google Gemini API key with Knit through the official OpenAI-compatible endpoint. Gemini is bring-your-own only and needs just the key.
Read guide →LLM provider
Ollama (local or cloud) to Knit
Run open models with Ollama and connect them to Knit. Ollama needs a publicly reachable endpoint — a tunnel for local instances, or Ollama Cloud with a Bearer key.
Read guide →Tools
an MCP server to Knit
Add Model Context Protocol (MCP) servers to Knit to give the AI extra tools. Knit supports Streamable HTTP MCP servers with an optional Authorization header.
Read guide →