Skip to content

AI Providers

Jump to bottom
Sasha Baranov edited this page Jun 12, 2026 · 1 revision

AI Providers

Aicraft uses the OpenAI chat completions format: POST /v1/chat/completions with a JSON body containing model and messages.

Ollama (recommended for local)

On the same machine as Minecraft:

ollama pull llama3.2
ollama serve
endpoint: "http://localhost:11434/v1/chat/completions"
api-key: "ollama"
model: "llama3.2"

If Ollama runs on another machine, replace localhost with that host's IP.

OpenAI

endpoint: "https://api.openai.com/v1/chat/completions"
api-key: "sk-your-key-here"
model: "gpt-4o-mini"

Other compatible providers

Any service exposing /v1/chat/completions works, including:

  • Unsloth
  • vLLM
  • LM Studio
  • Groq
  • OpenRouter

Set endpoint to the full completions URL, api-key to your key (or a placeholder if unused), and model to the provider's model name.

Docker note

If the AI server runs in Docker and Minecraft runs on the host, localhost from the Minecraft server's perspective may not reach the container. Use the container's published IP/port or host.docker.internal where applicable.

Clone this wiki locally