LM Studio Setup

Overview

LM Studio is a free desktop app for running open-source models locally — Llama, Qwen, GLM, Kimi, DeepSeek, gpt-oss, and anything else in GGUF or MLX format. Since v0.3.17 it supports MCP servers, including remote servers with OAuth.

Connecting Tempreon gives those local models the thing they otherwise lack: persistence. Your AI loads your Core Imprint, searches your Knowledge Vault, and remembers what it learns — across every tool. This is the cleanest way to give a fully local model a persistent identity — the model runs entirely on your machine, and Tempreon supplies the memory.

Quick path: Chat sidebar → Program tab → Install > Edit mcp.json → add the Tempreon entry → save → approve the OAuth prompt in your browser.

One-click install: with LM Studio installed, Add to LM Studio adds the entry for you — then approve the OAuth prompt in your browser.

Prerequisites

• LM Studio v0.3.17 or later (free), with at least one model downloaded
• A model that supports tool calling — check the model card; not every local model does
• A Tempreon account with at least one available Bridge slot
• Your Tempreon MCP server URL (also shown in your Bridges dashboard)

Step-by-Step Setup

1. Open the mcp.json editor

1. In a chat, open the right-hand sidebar
2. Click the Program tab
3. Click Install > Edit mcp.json

2. Add the Tempreon entry

Add Tempreon to the mcpServers object and save the file:

{
  "mcpServers": {
    "tempreon": {
      "url": "https://api.tempreon.com/functions/v1/tempreon-mcp/mcp"
    }
  }
}

3. Authorize in your browser

When you save, LM Studio opens your browser with the Tempreon consent screen automatically. Sign in if prompted and approve. Tempreon supports dynamic client registration, so there are no keys, client IDs, or secrets to paste.

4. Approve the first tool calls

LM Studio asks for confirmation before each tool call. When your model first calls a Tempreon tool (such as session_start), a confirmation dialog appears — approve it. The same dialog lets you relax future confirmations for tools you trust.

Troubleshooting

The model never calls Tempreon tools. Confirm your model supports tool calling — many local models don't, or do it unreliably. Pick a model whose card explicitly lists tool or function calling.

Responses got worse after connecting. Smaller local models can struggle when a large tool list lands in their context. Try a larger model, or disable the Tempreon tools you don't need for the current chat.

No browser window opens for authorization. Re-save mcp.json to retrigger the flow, and check that a default browser is set at the OS level.

The connection fails immediately. Confirm the URL ends in /mcp with no trailing slash: https://api.tempreon.com/functions/v1/tempreon-mcp/mcp.

Also connect: Open-Source Chat Clients · Raycast · Warp

Try saying...

• "Start my session." — triggers session_start to load your Core Imprint
• "What do you know about me from Tempreon?" — a quick end-to-end test
• "Remember that [insight] for next time." — proves your local model can write to your memory