LM Studio guide: connect local models in Silly Tavern and MiniTavern

LM Studio runs GGUF models on your desktop and exposes an OpenAI-compatible API via Local Server (often port 1234). This guide covers setup and wiring for Silly Tavern on PC and MiniTavern on your phone.

Step-by-step MiniTavern setup: official tutorial. If LM Studio “vanished” after ST 1.16+, see find Ollama / LM Studio in three steps.

---

Prepare LM Studio

1. Install from lmstudio.ai.
2. Download and load a model (e.g. a 7B instruct GGUF).
3. Start Local Server under Developer.
4. Test: curl http://127.0.0.1:1234/v1/models
5. For phone access: enable network connections and use your LAN IP.

---

Silly Tavern

1. API Connections → main API: Text Completion.
2. API Type: Generic (OpenAI-compatible) [LM Studio, LiteLLM, etc.].
3. Server URL: http://127.0.0.1:1234
4. API Key: empty or lm-studio.
5. Connect and pick the loaded model.

Setting	Example
Main API	Text Completion
API Type	Generic (OpenAI-compatible)
Server URL	http://127.0.0.1:1234

---

MiniTavern

1. Keep Local Server running on your PC (LAN allowed).
2. Other LLM → URL: http://192.168.1.8:1234/v1 (your PC IP).
3. API Key: lm-studio or any placeholder.
4. Refresh models, test, save.

Do not use Ollama’s 11434/api URL for LM Studio.

---

About the author

花（Hana）

AI工具評価の専門家。東京・新宿三丁目周辺で活動し、最新のAIアプリケーションやツールを実際に使用してレビューを提供しています。

@xluuuu92030

---

Published: June 5, 2026
Updated: June 5, 2026