LLM NPC Plugin — Lite

Give your Citizens NPCs real conversations powered by a local LLM — no cloud API, no per-message cost. Bind one NPC to a personality file, and players can just talk to it in chat.

Features

Features

• Local-first: runs against LM Studio on your own hardware — zero API cost, zero data leaving your server
• Prompt-injection hardening and startup config validation (same safety floor as the paid edition — never compromised in Lite)
• Streaming replies so players see responses arrive naturally
• Rate limiting and graceful fallback if the LLM is slow or down
• 3-turn conversation memory per player/NPC
• Full messages.yml localization

Requirements

• Citizens — required for NPC interaction
• A local OpenAI-compatible chat-completions server, running on your machine or network. Any of these work:
  • LM Studio — default config.yml endpoint (http://localhost:1234/v1/chat/completions) is preconfigured for this
  • Ollama — set llm.endpoint to http://localhost:11434/v1/chat/completions and llm.model to your pulled model name (e.g. qwen2.5:7b)
  • Any other server exposing an OpenAI-compatible /v1/chat/completions route (text-generation-webui, vLLM, llama.cpp server, etc.)
• No cloud API key needed, no data leaves your machine

Lite vs Premium

Lite caps you at 1 bound NPC and 3-turn memory to keep it genuinely free and sustainable. Need more NPCs, deeper memory, cloud provider support, the developer API, or PlaceholderAPI integration? Check out the Premium listing.